From 12270471e147de5616c4669727bbb1114a2e2abb Mon Sep 17 00:00:00 2001 From: Roland Tannous Date: Tue, 18 Nov 2025 06:21:35 +0000 Subject: [PATCH 01/12] fix TorchAOConfig' object has no attribute 'base_config' error --- nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb | 202 ++++++++++-------- nb/Qwen3_(4B)_Instruct-QAT.ipynb | 202 ++++++++++-------- .../Kaggle-Qwen3_(4B)_Instruct-QAT.py | 2 +- python_scripts/Qwen3_(4B)_Instruct-QAT.py | 2 +- 4 files changed, 220 insertions(+), 188 deletions(-) diff --git a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb index a519c29a..60d45f14 100644 --- a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", + " Join Discord if you need help + ⭐ Star us on Github ⭐\n", "
\n", "\n", "To install Unsloth your local device, follow [our guide](https://docs.unsloth.ai/get-started/install-and-update). This notebook is licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme).\n", @@ -51,7 +51,22 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" + "source": [ + "%%capture\n", + "import os, re\n", + "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", + " !pip install unsloth\n", + "else:\n", + " # Do this only in Colab notebooks! Otherwise use pip install unsloth\n", + " import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n", + " xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n", + " !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n", + " !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n", + " !pip install --no-deps unsloth\n", + "!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n", + "!pip install transformers==4.55.4\n", + "!pip install --no-deps trl==0.22.2" + ] }, { "cell_type": "markdown", @@ -212,7 +227,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "\ud83e\udda5 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n" + "πŸ¦₯ Unsloth: Will patch your computer to enable 2x faster free finetuning.\n" ] }, { @@ -227,7 +242,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "\ud83e\udda5 Unsloth Zoo will now patch everything to make training faster!\n", + "πŸ¦₯ Unsloth Zoo will now patch everything to make training faster!\n", "==((====))== Unsloth 2025.10.8: Fast Qwen3 patching. Transformers: 4.55.4.\n", " \\\\ /| Tesla T4. Num GPUs = 1. Max memory: 14.741 GB. Platform: Linux.\n", "O^O/ \\_/ \\ Torch: 2.8.0+cu126. CUDA: 7.5. CUDA Toolkit: 12.6. Triton: 3.4.0\n", @@ -1519,7 +1534,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Unsloth: Copying 2 files from cache to `model`: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [03:20<00:00, 100.23s/it]\n" + "Unsloth: Copying 2 files from cache to `model`: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [03:20<00:00, 100.23s/it]\n" ] }, { @@ -1536,8 +1551,8 @@ "name": "stderr", "output_type": "stream", "text": [ - "Unsloth: Preparing safetensor model files: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [00:00<00:00, 15709.00it/s]\n", - "Unsloth: Merging weights into 16bit: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [04:31<00:00, 135.70s/it]\n" + "Unsloth: Preparing safetensor model files: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [00:00<00:00, 15709.00it/s]\n", + "Unsloth: Merging weights into 16bit: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [04:31<00:00, 135.70s/it]\n" ] }, { @@ -1566,7 +1581,7 @@ "model.save_pretrained_torchao(\n", " \"model\",\n", " tokenizer,\n", - " torchao_config = model._torchao_config.base_config,\n", + " torchao_config = model._torchao_config.base_config_and_filter_fns[0][0],\n", ")" ] }, @@ -1639,7 +1654,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", + " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", "\n", " This notebook and all Unsloth notebooks are licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme)\n", "\n" @@ -1654,7 +1669,8 @@ "provenance": [] }, "kernelspec": { - "display_name": "Python 3", + "display_name": "Python 3 (ipykernel)", + "language": "python", "name": "python3" }, "language_info": { @@ -1667,7 +1683,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.10.10" + "version": "3.12.0" }, "widgets": { "application/vnd.jupyter.widget-state+json": { @@ -1687,9 +1703,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5471cb70b9fc476e924a79bae6fbe973", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_10ccee1030c94c05a299e4dbb646f26c", - "value": "\u2007100000/100000\u2007[00:52<00:00,\u20071750.73\u2007examples/s]" + "value": " 100000/100000 [00:52<00:00, 1750.73 examples/s]" } }, "0297fab8f99f4de6a905c5e16ecf6d45": { @@ -1782,9 +1798,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c168576851f240a9a6f0c9de5045ea76", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6f4d50b5f5db4d27a3219b6a83560542", - "value": "\u20079.65k/?\u2007[00:00<00:00,\u2007164kB/s]" + "value": " 9.65k/? [00:00<00:00, 164kB/s]" } }, "0658d6cd875e4ddfbcaec38acfc6ab8e": { @@ -2194,9 +2210,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d027fe1c1c874b00b35d1aa728e2f64b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c050f260067d4a8cb0a48b8789853e42", - "value": "model.safetensors.index.json:\u2007" + "value": "model.safetensors.index.json: " } }, "15493837f6ed4859a856749581e4bc94": { @@ -2283,9 +2299,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_eed8825725c8493aa839ee1f662981a8", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_873775127eff456e8418e310c10f68c6", - "value": "Generating\u2007train\u2007split:\u2007100%" + "value": "Generating train split: 100%" } }, "1aba00c5303e49d88c4bc8987c8e2b03": { @@ -2656,9 +2672,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1c977d30200d437abcc3bab5459ac3a0", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_edc64ecdba034ac5afc978bf8c844016", - "value": "\u20072/2\u2007[00:31<00:00,\u200714.00s/it]" + "value": " 2/2 [00:31<00:00, 14.00s/it]" } }, "23182fb1edd248259238b5eae55811e2": { @@ -2729,9 +2745,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c2986f10f6a64a2fa7bd4fdb70e08e04", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9efb3a9c377c4ce3bf13038cd8cb9c63", - "value": "\u2007100000/100000\u2007[00:04<00:00,\u200717940.57\u2007examples/s]" + "value": " 100000/100000 [00:04<00:00, 17940.57 examples/s]" } }, "24d4ad43783e4715b89a1bbd0a84b90c": { @@ -2811,9 +2827,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_74499d604e454b009f07f9900f010e4e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_3312b27851c64b50a3e52013b4285c13", - "value": "\u2007982/982\u2007[00:00<00:00,\u2007114kB/s]" + "value": " 982/982 [00:00<00:00, 114kB/s]" } }, "2a86e8d4b9d64fc990de342af2359327": { @@ -2936,9 +2952,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c8de0d52dd214fd283bf3d2ff9c50bd7", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a0f7475651e84e02afd306a2f6d8cb51", - "value": "Loading\u2007checkpoint\u2007shards:\u2007100%" + "value": "Loading checkpoint shards: 100%" } }, "36986d45f3f342f494287555aab6d79c": { @@ -3101,9 +3117,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_36986d45f3f342f494287555aab6d79c", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_7ad06c11c12e4d12bb8567f46e005940", - "value": "\u2007237/237\u2007[00:00<00:00,\u20076.90kB/s]" + "value": " 237/237 [00:00<00:00, 6.90kB/s]" } }, "3ff4a0687fdf44fa95afe9ec0ab781ea": { @@ -3174,9 +3190,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b7f2a83d93994f83b4163627a6619c72", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_47586e1fc8cb454e84d447d3852fb8fc", - "value": "\u20072/2\u2007[00:44<00:00,\u200721.28s/it]" + "value": " 2/2 [00:44<00:00, 21.28s/it]" } }, "40b609ebe3c2437a9ec895e48de6d1ca": { @@ -3351,9 +3367,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4b78a2ab29ba40bb98e18bf3beb1181e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5d6d1d2d76dd45d4a6ec9d1cc067968c", - "value": "model-00002-of-00002.safetensors:\u2007100%" + "value": "model-00002-of-00002.safetensors: 100%" } }, "4641a1c322e94997a1d4e99ba4d29db4": { @@ -3424,9 +3440,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_54bb3cb7e89843e881366b277107f10e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_b48febda135143208ab19ddbc54db83f", - "value": "\u20074.04k/?\u2007[00:00<00:00,\u2007405kB/s]" + "value": " 4.04k/? [00:00<00:00, 405kB/s]" } }, "47586e1fc8cb454e84d447d3852fb8fc": { @@ -3550,9 +3566,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ea93bb20bbc44092ab0b1394cbd28d4e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_06d793c0fe3f46e29e279674de222b8f", - "value": "\u2007100000/100000\u2007[04:18<00:00,\u2007464.70\u2007examples/s]" + "value": " 100000/100000 [04:18<00:00, 464.70 examples/s]" } }, "4d4f8726c88d40bcabdb664a663cbcc8": { @@ -3654,9 +3670,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c88063243cb0423db943d5c14b53b799", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_af75f290b4e1431681085b82b894129c", - "value": "\u20073.08G/3.08G\u2007[00:34<00:00,\u2007244MB/s]" + "value": " 3.08G/3.08G [00:34<00:00, 244MB/s]" } }, "5471cb70b9fc476e924a79bae6fbe973": { @@ -3855,9 +3871,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_63a2c8a60c494acd87b00b4cfdff5c68", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_adbddacbb5414d24a37c1a4a0f036547", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "564ec00e94eb477785335399eb492556": { @@ -3876,9 +3892,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1aba00c5303e49d88c4bc8987c8e2b03", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_bf150441841847489593ab9cffc385ac", - "value": "chat_template.jinja:\u2007" + "value": "chat_template.jinja: " } }, "573ad1a363b746699575ad8d77592312": { @@ -4436,9 +4452,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fc3356580e6744ca83b8649368480b3e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5809cbdec1844cbebc0db710185d8072", - "value": "Unsloth:\u2007Standardizing\u2007formats\u2007(num_proc=2):\u2007100%" + "value": "Unsloth: Standardizing formats (num_proc=2): 100%" } }, "63a2c8a60c494acd87b00b4cfdff5c68": { @@ -4577,9 +4593,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_f72b1f631cc6458dbbdbc4fc3642133f", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c0a0d7c3181442c69fd330ed1452dd84", - "value": "merges.txt:\u2007" + "value": "merges.txt: " } }, "67c3227617ea4f72a5a3a9e822bb8abb": { @@ -5130,9 +5146,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a1cba3ed46c34f2ebfaf2b1dbb72d0d3", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_573ad1a363b746699575ad8d77592312", - "value": "vocab.json:\u2007" + "value": "vocab.json: " } }, "803af90c303241edbbca3338dd4f043d": { @@ -5218,9 +5234,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_613103c611fd412c815a37de1007dd3d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_78a08d6e960f43ad9b74f46c125b5aea", - "value": "data/train-00000-of-00001.parquet:\u2007100%" + "value": "data/train-00000-of-00001.parquet: 100%" } }, "854ab752c85449ad80b68ad7677df198": { @@ -5291,9 +5307,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_598cdec8e8dd49a0bec21b74394ad47d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_da70610720a14a1ca0e689850a081809", - "value": "tokenizer_config.json:\u2007" + "value": "tokenizer_config.json: " } }, "8e0aa0bf9fdf456bae3ff6dae9825eac": { @@ -5456,9 +5472,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6226ba2708f54d3682c2257422299b61", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5b790390fed14b339b58be310eaf5b7b", - "value": "\u2007614/614\u2007[00:00<00:00,\u200713.3kB/s]" + "value": " 614/614 [00:00<00:00, 13.3kB/s]" } }, "94561bbb46cc4636ab21428a8a47f107": { @@ -5492,9 +5508,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_40b609ebe3c2437a9ec895e48de6d1ca", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6e7a0eb4f9e340bd9883bc975c6a489c", - "value": "\u2007707/707\u2007[00:00<00:00,\u200710.8kB/s]" + "value": " 707/707 [00:00<00:00, 10.8kB/s]" } }, "984b8bee3e6e493e8bebaa177e9a8aa3": { @@ -5581,9 +5597,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ba99aaf582914c83b18910ac713a9b6f", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_94561bbb46cc4636ab21428a8a47f107", - "value": "\u2007100000/100000\u2007[00:32<00:00,\u20072848.96\u2007examples/s]" + "value": " 100000/100000 [00:32<00:00, 2848.96 examples/s]" } }, "9a89009072e647b2b760688428c76e06": { @@ -5730,9 +5746,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fa846c3ef9124e6090530fd0ad72851e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_3ab7781489bc4bbe9b072fd49f7c12f8", - "value": "Map\u2007(num_proc=6):\u2007100%" + "value": "Map (num_proc=6): 100%" } }, "9efb3a9c377c4ce3bf13038cd8cb9c63": { @@ -5983,9 +5999,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_61c268a7d13a44afa56bedade90e4f34", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_272afad7361c4eeab631c7bd41e6fa14", - "value": "\u20074.97G/4.97G\u2007[00:57<00:00,\u2007263MB/s]" + "value": " 4.97G/4.97G [00:57<00:00, 263MB/s]" } }, "a893ae8f4f954f78926d91a2a2b019de": { @@ -6072,9 +6088,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0e31ade8b833471a81309de8c821bce9", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_88b7cea7ba0e42b7866bbfa4dc2df056", - "value": "\u2007100000/100000\u2007[00:01<00:00,\u200797725.37\u2007examples/s]" + "value": " 100000/100000 [00:01<00:00, 97725.37 examples/s]" } }, "a95fca80392b4af3aae5d189059dace6": { @@ -6093,9 +6109,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4641a1c322e94997a1d4e99ba4d29db4", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e2e1b4e39b7c44caaf698082cef96f20", - "value": "Unsloth:\u2007Tokenizing\u2007["text"]\u2007(num_proc=6):\u2007100%" + "value": "Unsloth: Tokenizing ["text"] (num_proc=6): 100%" } }, "a98adcf1d23547f599e01ab83bec8b48": { @@ -6234,9 +6250,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c993d2d141094cf0b7151a8dbffd9a45", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_3d9701a29b684fb0ac32535accf25ccb", - "value": "\u20072.78M/?\u2007[00:00<00:00,\u200712.3MB/s]" + "value": " 2.78M/? [00:00<00:00, 12.3MB/s]" } }, "ac2ceaa6d5fe41169557e673ccc83ec5": { @@ -6359,9 +6375,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_cc8af4f890a54538ba6aae19866dc7c0", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_803af90c303241edbbca3338dd4f043d", - "value": "\u200711.4M/11.4M\u2007[00:01<00:00,\u20079.26MB/s]" + "value": " 11.4M/11.4M [00:01<00:00, 9.26MB/s]" } }, "affc3e6798984ab7bee3b100e9bf07f3": { @@ -6380,9 +6396,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_79ac14919f7c47de914f634481c7c47a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_611831076b59469fbcfcacdec4ac4bea", - "value": "\u200732.9k/?\u2007[00:00<00:00,\u20073.04MB/s]" + "value": " 32.9k/? [00:00<00:00, 3.04MB/s]" } }, "b021cc01ce3748d48aa7fc32843c51f9": { @@ -6942,9 +6958,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1ea6c960f5d3449eb445ba26b226ab80", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_b7a3d157a73644c899f1e4d60678d9c3", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "c5be281628bd49f0b184cd899143c2c4": { @@ -7410,9 +7426,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_592370835995436e8ce573f9f121b410", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_24d4ad43783e4715b89a1bbd0a84b90c", - "value": "Map:\u2007100%" + "value": "Map: 100%" } }, "d38043a7e0ce4eeabd900a7ad01fb733": { @@ -7453,9 +7469,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a6c5b152ce8a4845aab8f4bd42bbc081", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_2a86e8d4b9d64fc990de342af2359327", - "value": "Loading\u2007checkpoint\u2007shards:\u2007100%" + "value": "Loading checkpoint shards: 100%" } }, "d5ac21ecf53146b9a91e77c1365a8087": { @@ -7633,9 +7649,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a8f346bd2311473bb41f9c832d38fde3", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_27080c04e0f44a21986825766793e12a", - "value": "generation_config.json:\u2007100%" + "value": "generation_config.json: 100%" } }, "dfe5475d5906460e9a771e0eeefdcd46": { @@ -7654,9 +7670,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_7648f19accea41f5abec2f3d8296ac19", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a34d8312550e4c89a37dafb88a906dcd", - "value": "\u2007117M/117M\u2007[00:01<00:00,\u2007118MB/s]" + "value": " 117M/117M [00:01<00:00, 118MB/s]" } }, "e01669badd9043e998951c212c8ff174": { @@ -7864,9 +7880,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6eecd84361254cf6973f16e70265707b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0fa3c3feb88042afbd921a51c91353ee", - "value": "added_tokens.json:\u2007100%" + "value": "added_tokens.json: 100%" } }, "e86016cd0c124ca78fe3119e59c94de8": { @@ -7907,9 +7923,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_3ff4a0687fdf44fa95afe9ec0ab781ea", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a461d5efcb48465eb6c1b39a7d67fa07", - "value": "README.md:\u2007100%" + "value": "README.md: 100%" } }, "ea888f9b76b448208a11d6eee0c1eb81": { @@ -8169,9 +8185,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c5be281628bd49f0b184cd899143c2c4", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_4d4f8726c88d40bcabdb664a663cbcc8", - "value": "model-00001-of-00002.safetensors:\u2007100%" + "value": "model-00001-of-00002.safetensors: 100%" } }, "f5279e94f100416e858be5122226ef53": { @@ -8264,9 +8280,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_577decf0054b409a95ff0a0bf23b73ed", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_cce32a3c7860476690a0e5762f22b1ab", - "value": "\u20071.67M/?\u2007[00:00<00:00,\u20077.39MB/s]" + "value": " 1.67M/? [00:00<00:00, 7.39MB/s]" } }, "f81c5ed93c684019b289c7a01ec0f177": { @@ -8516,5 +8532,5 @@ } }, "nbformat": 4, - "nbformat_minor": 0 -} \ No newline at end of file + "nbformat_minor": 4 +} diff --git a/nb/Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Qwen3_(4B)_Instruct-QAT.ipynb index a519c29a..60d45f14 100644 --- a/nb/Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Qwen3_(4B)_Instruct-QAT.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", + " Join Discord if you need help + ⭐ Star us on Github ⭐\n", "
\n", "\n", "To install Unsloth your local device, follow [our guide](https://docs.unsloth.ai/get-started/install-and-update). This notebook is licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme).\n", @@ -51,7 +51,22 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" + "source": [ + "%%capture\n", + "import os, re\n", + "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", + " !pip install unsloth\n", + "else:\n", + " # Do this only in Colab notebooks! Otherwise use pip install unsloth\n", + " import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n", + " xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n", + " !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n", + " !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n", + " !pip install --no-deps unsloth\n", + "!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n", + "!pip install transformers==4.55.4\n", + "!pip install --no-deps trl==0.22.2" + ] }, { "cell_type": "markdown", @@ -212,7 +227,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "\ud83e\udda5 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n" + "πŸ¦₯ Unsloth: Will patch your computer to enable 2x faster free finetuning.\n" ] }, { @@ -227,7 +242,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "\ud83e\udda5 Unsloth Zoo will now patch everything to make training faster!\n", + "πŸ¦₯ Unsloth Zoo will now patch everything to make training faster!\n", "==((====))== Unsloth 2025.10.8: Fast Qwen3 patching. Transformers: 4.55.4.\n", " \\\\ /| Tesla T4. Num GPUs = 1. Max memory: 14.741 GB. Platform: Linux.\n", "O^O/ \\_/ \\ Torch: 2.8.0+cu126. CUDA: 7.5. CUDA Toolkit: 12.6. Triton: 3.4.0\n", @@ -1519,7 +1534,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Unsloth: Copying 2 files from cache to `model`: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [03:20<00:00, 100.23s/it]\n" + "Unsloth: Copying 2 files from cache to `model`: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [03:20<00:00, 100.23s/it]\n" ] }, { @@ -1536,8 +1551,8 @@ "name": "stderr", "output_type": "stream", "text": [ - "Unsloth: Preparing safetensor model files: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [00:00<00:00, 15709.00it/s]\n", - "Unsloth: Merging weights into 16bit: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [04:31<00:00, 135.70s/it]\n" + "Unsloth: Preparing safetensor model files: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [00:00<00:00, 15709.00it/s]\n", + "Unsloth: Merging weights into 16bit: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [04:31<00:00, 135.70s/it]\n" ] }, { @@ -1566,7 +1581,7 @@ "model.save_pretrained_torchao(\n", " \"model\",\n", " tokenizer,\n", - " torchao_config = model._torchao_config.base_config,\n", + " torchao_config = model._torchao_config.base_config_and_filter_fns[0][0],\n", ")" ] }, @@ -1639,7 +1654,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", + " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", "\n", " This notebook and all Unsloth notebooks are licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme)\n", "\n" @@ -1654,7 +1669,8 @@ "provenance": [] }, "kernelspec": { - "display_name": "Python 3", + "display_name": "Python 3 (ipykernel)", + "language": "python", "name": "python3" }, "language_info": { @@ -1667,7 +1683,7 @@ "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", - "version": "3.10.10" + "version": "3.12.0" }, "widgets": { "application/vnd.jupyter.widget-state+json": { @@ -1687,9 +1703,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5471cb70b9fc476e924a79bae6fbe973", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_10ccee1030c94c05a299e4dbb646f26c", - "value": "\u2007100000/100000\u2007[00:52<00:00,\u20071750.73\u2007examples/s]" + "value": " 100000/100000 [00:52<00:00, 1750.73 examples/s]" } }, "0297fab8f99f4de6a905c5e16ecf6d45": { @@ -1782,9 +1798,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c168576851f240a9a6f0c9de5045ea76", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6f4d50b5f5db4d27a3219b6a83560542", - "value": "\u20079.65k/?\u2007[00:00<00:00,\u2007164kB/s]" + "value": " 9.65k/? [00:00<00:00, 164kB/s]" } }, "0658d6cd875e4ddfbcaec38acfc6ab8e": { @@ -2194,9 +2210,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d027fe1c1c874b00b35d1aa728e2f64b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c050f260067d4a8cb0a48b8789853e42", - "value": "model.safetensors.index.json:\u2007" + "value": "model.safetensors.index.json: " } }, "15493837f6ed4859a856749581e4bc94": { @@ -2283,9 +2299,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_eed8825725c8493aa839ee1f662981a8", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_873775127eff456e8418e310c10f68c6", - "value": "Generating\u2007train\u2007split:\u2007100%" + "value": "Generating train split: 100%" } }, "1aba00c5303e49d88c4bc8987c8e2b03": { @@ -2656,9 +2672,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1c977d30200d437abcc3bab5459ac3a0", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_edc64ecdba034ac5afc978bf8c844016", - "value": "\u20072/2\u2007[00:31<00:00,\u200714.00s/it]" + "value": " 2/2 [00:31<00:00, 14.00s/it]" } }, "23182fb1edd248259238b5eae55811e2": { @@ -2729,9 +2745,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c2986f10f6a64a2fa7bd4fdb70e08e04", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_9efb3a9c377c4ce3bf13038cd8cb9c63", - "value": "\u2007100000/100000\u2007[00:04<00:00,\u200717940.57\u2007examples/s]" + "value": " 100000/100000 [00:04<00:00, 17940.57 examples/s]" } }, "24d4ad43783e4715b89a1bbd0a84b90c": { @@ -2811,9 +2827,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_74499d604e454b009f07f9900f010e4e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_3312b27851c64b50a3e52013b4285c13", - "value": "\u2007982/982\u2007[00:00<00:00,\u2007114kB/s]" + "value": " 982/982 [00:00<00:00, 114kB/s]" } }, "2a86e8d4b9d64fc990de342af2359327": { @@ -2936,9 +2952,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c8de0d52dd214fd283bf3d2ff9c50bd7", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a0f7475651e84e02afd306a2f6d8cb51", - "value": "Loading\u2007checkpoint\u2007shards:\u2007100%" + "value": "Loading checkpoint shards: 100%" } }, "36986d45f3f342f494287555aab6d79c": { @@ -3101,9 +3117,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_36986d45f3f342f494287555aab6d79c", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_7ad06c11c12e4d12bb8567f46e005940", - "value": "\u2007237/237\u2007[00:00<00:00,\u20076.90kB/s]" + "value": " 237/237 [00:00<00:00, 6.90kB/s]" } }, "3ff4a0687fdf44fa95afe9ec0ab781ea": { @@ -3174,9 +3190,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b7f2a83d93994f83b4163627a6619c72", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_47586e1fc8cb454e84d447d3852fb8fc", - "value": "\u20072/2\u2007[00:44<00:00,\u200721.28s/it]" + "value": " 2/2 [00:44<00:00, 21.28s/it]" } }, "40b609ebe3c2437a9ec895e48de6d1ca": { @@ -3351,9 +3367,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4b78a2ab29ba40bb98e18bf3beb1181e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5d6d1d2d76dd45d4a6ec9d1cc067968c", - "value": "model-00002-of-00002.safetensors:\u2007100%" + "value": "model-00002-of-00002.safetensors: 100%" } }, "4641a1c322e94997a1d4e99ba4d29db4": { @@ -3424,9 +3440,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_54bb3cb7e89843e881366b277107f10e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_b48febda135143208ab19ddbc54db83f", - "value": "\u20074.04k/?\u2007[00:00<00:00,\u2007405kB/s]" + "value": " 4.04k/? [00:00<00:00, 405kB/s]" } }, "47586e1fc8cb454e84d447d3852fb8fc": { @@ -3550,9 +3566,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ea93bb20bbc44092ab0b1394cbd28d4e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_06d793c0fe3f46e29e279674de222b8f", - "value": "\u2007100000/100000\u2007[04:18<00:00,\u2007464.70\u2007examples/s]" + "value": " 100000/100000 [04:18<00:00, 464.70 examples/s]" } }, "4d4f8726c88d40bcabdb664a663cbcc8": { @@ -3654,9 +3670,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c88063243cb0423db943d5c14b53b799", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_af75f290b4e1431681085b82b894129c", - "value": "\u20073.08G/3.08G\u2007[00:34<00:00,\u2007244MB/s]" + "value": " 3.08G/3.08G [00:34<00:00, 244MB/s]" } }, "5471cb70b9fc476e924a79bae6fbe973": { @@ -3855,9 +3871,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_63a2c8a60c494acd87b00b4cfdff5c68", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_adbddacbb5414d24a37c1a4a0f036547", - "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json: 100%" } }, "564ec00e94eb477785335399eb492556": { @@ -3876,9 +3892,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1aba00c5303e49d88c4bc8987c8e2b03", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_bf150441841847489593ab9cffc385ac", - "value": "chat_template.jinja:\u2007" + "value": "chat_template.jinja: " } }, "573ad1a363b746699575ad8d77592312": { @@ -4436,9 +4452,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fc3356580e6744ca83b8649368480b3e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5809cbdec1844cbebc0db710185d8072", - "value": "Unsloth:\u2007Standardizing\u2007formats\u2007(num_proc=2):\u2007100%" + "value": "Unsloth: Standardizing formats (num_proc=2): 100%" } }, "63a2c8a60c494acd87b00b4cfdff5c68": { @@ -4577,9 +4593,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_f72b1f631cc6458dbbdbc4fc3642133f", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_c0a0d7c3181442c69fd330ed1452dd84", - "value": "merges.txt:\u2007" + "value": "merges.txt: " } }, "67c3227617ea4f72a5a3a9e822bb8abb": { @@ -5130,9 +5146,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a1cba3ed46c34f2ebfaf2b1dbb72d0d3", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_573ad1a363b746699575ad8d77592312", - "value": "vocab.json:\u2007" + "value": "vocab.json: " } }, "803af90c303241edbbca3338dd4f043d": { @@ -5218,9 +5234,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_613103c611fd412c815a37de1007dd3d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_78a08d6e960f43ad9b74f46c125b5aea", - "value": "data/train-00000-of-00001.parquet:\u2007100%" + "value": "data/train-00000-of-00001.parquet: 100%" } }, "854ab752c85449ad80b68ad7677df198": { @@ -5291,9 +5307,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_598cdec8e8dd49a0bec21b74394ad47d", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_da70610720a14a1ca0e689850a081809", - "value": "tokenizer_config.json:\u2007" + "value": "tokenizer_config.json: " } }, "8e0aa0bf9fdf456bae3ff6dae9825eac": { @@ -5456,9 +5472,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6226ba2708f54d3682c2257422299b61", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_5b790390fed14b339b58be310eaf5b7b", - "value": "\u2007614/614\u2007[00:00<00:00,\u200713.3kB/s]" + "value": " 614/614 [00:00<00:00, 13.3kB/s]" } }, "94561bbb46cc4636ab21428a8a47f107": { @@ -5492,9 +5508,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_40b609ebe3c2437a9ec895e48de6d1ca", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_6e7a0eb4f9e340bd9883bc975c6a489c", - "value": "\u2007707/707\u2007[00:00<00:00,\u200710.8kB/s]" + "value": " 707/707 [00:00<00:00, 10.8kB/s]" } }, "984b8bee3e6e493e8bebaa177e9a8aa3": { @@ -5581,9 +5597,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ba99aaf582914c83b18910ac713a9b6f", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_94561bbb46cc4636ab21428a8a47f107", - "value": "\u2007100000/100000\u2007[00:32<00:00,\u20072848.96\u2007examples/s]" + "value": " 100000/100000 [00:32<00:00, 2848.96 examples/s]" } }, "9a89009072e647b2b760688428c76e06": { @@ -5730,9 +5746,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fa846c3ef9124e6090530fd0ad72851e", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_3ab7781489bc4bbe9b072fd49f7c12f8", - "value": "Map\u2007(num_proc=6):\u2007100%" + "value": "Map (num_proc=6): 100%" } }, "9efb3a9c377c4ce3bf13038cd8cb9c63": { @@ -5983,9 +5999,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_61c268a7d13a44afa56bedade90e4f34", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_272afad7361c4eeab631c7bd41e6fa14", - "value": "\u20074.97G/4.97G\u2007[00:57<00:00,\u2007263MB/s]" + "value": " 4.97G/4.97G [00:57<00:00, 263MB/s]" } }, "a893ae8f4f954f78926d91a2a2b019de": { @@ -6072,9 +6088,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0e31ade8b833471a81309de8c821bce9", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_88b7cea7ba0e42b7866bbfa4dc2df056", - "value": "\u2007100000/100000\u2007[00:01<00:00,\u200797725.37\u2007examples/s]" + "value": " 100000/100000 [00:01<00:00, 97725.37 examples/s]" } }, "a95fca80392b4af3aae5d189059dace6": { @@ -6093,9 +6109,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4641a1c322e94997a1d4e99ba4d29db4", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_e2e1b4e39b7c44caaf698082cef96f20", - "value": "Unsloth:\u2007Tokenizing\u2007["text"]\u2007(num_proc=6):\u2007100%" + "value": "Unsloth: Tokenizing ["text"] (num_proc=6): 100%" } }, "a98adcf1d23547f599e01ab83bec8b48": { @@ -6234,9 +6250,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c993d2d141094cf0b7151a8dbffd9a45", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_3d9701a29b684fb0ac32535accf25ccb", - "value": "\u20072.78M/?\u2007[00:00<00:00,\u200712.3MB/s]" + "value": " 2.78M/? [00:00<00:00, 12.3MB/s]" } }, "ac2ceaa6d5fe41169557e673ccc83ec5": { @@ -6359,9 +6375,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_cc8af4f890a54538ba6aae19866dc7c0", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_803af90c303241edbbca3338dd4f043d", - "value": "\u200711.4M/11.4M\u2007[00:01<00:00,\u20079.26MB/s]" + "value": " 11.4M/11.4M [00:01<00:00, 9.26MB/s]" } }, "affc3e6798984ab7bee3b100e9bf07f3": { @@ -6380,9 +6396,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_79ac14919f7c47de914f634481c7c47a", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_611831076b59469fbcfcacdec4ac4bea", - "value": "\u200732.9k/?\u2007[00:00<00:00,\u20073.04MB/s]" + "value": " 32.9k/? [00:00<00:00, 3.04MB/s]" } }, "b021cc01ce3748d48aa7fc32843c51f9": { @@ -6942,9 +6958,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1ea6c960f5d3449eb445ba26b226ab80", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_b7a3d157a73644c899f1e4d60678d9c3", - "value": "special_tokens_map.json:\u2007100%" + "value": "special_tokens_map.json: 100%" } }, "c5be281628bd49f0b184cd899143c2c4": { @@ -7410,9 +7426,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_592370835995436e8ce573f9f121b410", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_24d4ad43783e4715b89a1bbd0a84b90c", - "value": "Map:\u2007100%" + "value": "Map: 100%" } }, "d38043a7e0ce4eeabd900a7ad01fb733": { @@ -7453,9 +7469,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a6c5b152ce8a4845aab8f4bd42bbc081", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_2a86e8d4b9d64fc990de342af2359327", - "value": "Loading\u2007checkpoint\u2007shards:\u2007100%" + "value": "Loading checkpoint shards: 100%" } }, "d5ac21ecf53146b9a91e77c1365a8087": { @@ -7633,9 +7649,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a8f346bd2311473bb41f9c832d38fde3", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_27080c04e0f44a21986825766793e12a", - "value": "generation_config.json:\u2007100%" + "value": "generation_config.json: 100%" } }, "dfe5475d5906460e9a771e0eeefdcd46": { @@ -7654,9 +7670,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_7648f19accea41f5abec2f3d8296ac19", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a34d8312550e4c89a37dafb88a906dcd", - "value": "\u2007117M/117M\u2007[00:01<00:00,\u2007118MB/s]" + "value": " 117M/117M [00:01<00:00, 118MB/s]" } }, "e01669badd9043e998951c212c8ff174": { @@ -7864,9 +7880,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6eecd84361254cf6973f16e70265707b", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_0fa3c3feb88042afbd921a51c91353ee", - "value": "added_tokens.json:\u2007100%" + "value": "added_tokens.json: 100%" } }, "e86016cd0c124ca78fe3119e59c94de8": { @@ -7907,9 +7923,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_3ff4a0687fdf44fa95afe9ec0ab781ea", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_a461d5efcb48465eb6c1b39a7d67fa07", - "value": "README.md:\u2007100%" + "value": "README.md: 100%" } }, "ea888f9b76b448208a11d6eee0c1eb81": { @@ -8169,9 +8185,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c5be281628bd49f0b184cd899143c2c4", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_4d4f8726c88d40bcabdb664a663cbcc8", - "value": "model-00001-of-00002.safetensors:\u2007100%" + "value": "model-00001-of-00002.safetensors: 100%" } }, "f5279e94f100416e858be5122226ef53": { @@ -8264,9 +8280,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_577decf0054b409a95ff0a0bf23b73ed", - "placeholder": "\u200b", + "placeholder": "​", "style": "IPY_MODEL_cce32a3c7860476690a0e5762f22b1ab", - "value": "\u20071.67M/?\u2007[00:00<00:00,\u20077.39MB/s]" + "value": " 1.67M/? [00:00<00:00, 7.39MB/s]" } }, "f81c5ed93c684019b289c7a01ec0f177": { @@ -8516,5 +8532,5 @@ } }, "nbformat": 4, - "nbformat_minor": 0 -} \ No newline at end of file + "nbformat_minor": 4 +} diff --git a/python_scripts/Kaggle-Qwen3_(4B)_Instruct-QAT.py b/python_scripts/Kaggle-Qwen3_(4B)_Instruct-QAT.py index 34529575..6d0949c9 100644 --- a/python_scripts/Kaggle-Qwen3_(4B)_Instruct-QAT.py +++ b/python_scripts/Kaggle-Qwen3_(4B)_Instruct-QAT.py @@ -343,7 +343,7 @@ def formatting_prompts_func(examples): model.save_pretrained_torchao( "model", tokenizer, - torchao_config = model._torchao_config.base_config, + torchao_config = model._torchao_config.base_config_and_filter_fns[0][0], ) diff --git a/python_scripts/Qwen3_(4B)_Instruct-QAT.py b/python_scripts/Qwen3_(4B)_Instruct-QAT.py index 34529575..6d0949c9 100644 --- a/python_scripts/Qwen3_(4B)_Instruct-QAT.py +++ b/python_scripts/Qwen3_(4B)_Instruct-QAT.py @@ -343,7 +343,7 @@ def formatting_prompts_func(examples): model.save_pretrained_torchao( "model", tokenizer, - torchao_config = model._torchao_config.base_config, + torchao_config = model._torchao_config.base_config_and_filter_fns[0][0], ) From 836fc5b3c4d567ebbfcab6dbb7fe0e063bcea11b Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 09:55:29 +0200 Subject: [PATCH 02/12] Update Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb revert unicode characters --- nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb | 177 ++++++++++++------------ 1 file changed, 88 insertions(+), 89 deletions(-) diff --git a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb index 60d45f14..64956a9b 100644 --- a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + ⭐ Star us on Github ⭐\n", + " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", "
\n", "\n", "To install Unsloth your local device, follow [our guide](https://docs.unsloth.ai/get-started/install-and-update). This notebook is licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme).\n", @@ -227,7 +227,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "πŸ¦₯ Unsloth: Will patch your computer to enable 2x faster free finetuning.\n" + "\ud83e\udda5 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n" ] }, { @@ -242,7 +242,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "πŸ¦₯ Unsloth Zoo will now patch everything to make training faster!\n", + "\ud83e\udda5 Unsloth Zoo will now patch everything to make training faster!\n", "==((====))== Unsloth 2025.10.8: Fast Qwen3 patching. Transformers: 4.55.4.\n", " \\\\ /| Tesla T4. Num GPUs = 1. Max memory: 14.741 GB. Platform: Linux.\n", "O^O/ \\_/ \\ Torch: 2.8.0+cu126. CUDA: 7.5. CUDA Toolkit: 12.6. Triton: 3.4.0\n", @@ -1534,7 +1534,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Unsloth: Copying 2 files from cache to `model`: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [03:20<00:00, 100.23s/it]\n" + "Unsloth: Copying 2 files from cache to `model`: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [03:20<00:00, 100.23s/it]\n" ] }, { @@ -1551,8 +1551,8 @@ "name": "stderr", "output_type": "stream", "text": [ - "Unsloth: Preparing safetensor model files: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [00:00<00:00, 15709.00it/s]\n", - "Unsloth: Merging weights into 16bit: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [04:31<00:00, 135.70s/it]\n" + "Unsloth: Preparing safetensor model files: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [00:00<00:00, 15709.00it/s]\n", + "Unsloth: Merging weights into 16bit: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [04:31<00:00, 135.70s/it]\n" ] }, { @@ -1654,7 +1654,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", + " Join Discord if you need help + \u2b50\ufe0f Star us on Github ⭐️\n", "\n", " This notebook and all Unsloth notebooks are licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme)\n", "\n" @@ -1669,8 +1669,7 @@ "provenance": [] }, "kernelspec": { - "display_name": "Python 3 (ipykernel)", - "language": "python", + "display_name": "Python 3", "name": "python3" }, "language_info": { @@ -1703,9 +1702,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5471cb70b9fc476e924a79bae6fbe973", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_10ccee1030c94c05a299e4dbb646f26c", - "value": " 100000/100000 [00:52<00:00, 1750.73 examples/s]" + "value": "\u2007100000/100000\u2007[00:52<00:00,\u20071750.73\u2007examples/s]" } }, "0297fab8f99f4de6a905c5e16ecf6d45": { @@ -1798,9 +1797,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c168576851f240a9a6f0c9de5045ea76", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_6f4d50b5f5db4d27a3219b6a83560542", - "value": " 9.65k/? [00:00<00:00, 164kB/s]" + "value": "\u20079.65k/?\u2007[00:00<00:00,\u2007164kB/s]" } }, "0658d6cd875e4ddfbcaec38acfc6ab8e": { @@ -2210,9 +2209,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d027fe1c1c874b00b35d1aa728e2f64b", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_c050f260067d4a8cb0a48b8789853e42", - "value": "model.safetensors.index.json: " + "value": "model.safetensors.index.json:\u2007" } }, "15493837f6ed4859a856749581e4bc94": { @@ -2299,9 +2298,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_eed8825725c8493aa839ee1f662981a8", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_873775127eff456e8418e310c10f68c6", - "value": "Generating train split: 100%" + "value": "Generating\u2007train\u2007split:\u2007100%" } }, "1aba00c5303e49d88c4bc8987c8e2b03": { @@ -2672,9 +2671,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1c977d30200d437abcc3bab5459ac3a0", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_edc64ecdba034ac5afc978bf8c844016", - "value": " 2/2 [00:31<00:00, 14.00s/it]" + "\u20072/2\u2007[00:31<00:00,\u200714.00s/it]" } }, "23182fb1edd248259238b5eae55811e2": { @@ -2745,9 +2744,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c2986f10f6a64a2fa7bd4fdb70e08e04", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_9efb3a9c377c4ce3bf13038cd8cb9c63", - "value": " 100000/100000 [00:04<00:00, 17940.57 examples/s]" + "value": "\u2007100000/100000\u2007[00:04<00:00,\u200717940.57\u2007examples/s]" } }, "24d4ad43783e4715b89a1bbd0a84b90c": { @@ -2827,9 +2826,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_74499d604e454b009f07f9900f010e4e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_3312b27851c64b50a3e52013b4285c13", - "value": " 982/982 [00:00<00:00, 114kB/s]" + "value": "\u2007982/982\u2007[00:00<00:00,\u2007114kB/s]" } }, "2a86e8d4b9d64fc990de342af2359327": { @@ -2952,9 +2951,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c8de0d52dd214fd283bf3d2ff9c50bd7", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_a0f7475651e84e02afd306a2f6d8cb51", - "value": "Loading checkpoint shards: 100%" + "value": "Loading\u2007checkpoint\u2007shards:\u2007100%" } }, "36986d45f3f342f494287555aab6d79c": { @@ -3117,9 +3116,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_36986d45f3f342f494287555aab6d79c", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_7ad06c11c12e4d12bb8567f46e005940", - "value": " 237/237 [00:00<00:00, 6.90kB/s]" + "value": "\u2007237/237\u2007[00:00<00:00,\u20076.90kB/s]" } }, "3ff4a0687fdf44fa95afe9ec0ab781ea": { @@ -3190,9 +3189,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b7f2a83d93994f83b4163627a6619c72", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_47586e1fc8cb454e84d447d3852fb8fc", - "value": " 2/2 [00:44<00:00, 21.28s/it]" + "\u20072/2\u2007[00:44<00:00,\u200721.28s/it]" } }, "40b609ebe3c2437a9ec895e48de6d1ca": { @@ -3367,9 +3366,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4b78a2ab29ba40bb98e18bf3beb1181e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_5d6d1d2d76dd45d4a6ec9d1cc067968c", - "value": "model-00002-of-00002.safetensors: 100%" + "value": "model-00002-of-00002.safetensors:\u2007100%" } }, "4641a1c322e94997a1d4e99ba4d29db4": { @@ -3440,9 +3439,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_54bb3cb7e89843e881366b277107f10e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_b48febda135143208ab19ddbc54db83f", - "value": " 4.04k/? [00:00<00:00, 405kB/s]" + "value": "value": "\u20074.04k/?\u2007[00:00<00:00,\u2007405kB/s]" } }, "47586e1fc8cb454e84d447d3852fb8fc": { @@ -3566,9 +3565,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ea93bb20bbc44092ab0b1394cbd28d4e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_06d793c0fe3f46e29e279674de222b8f", - "value": " 100000/100000 [04:18<00:00, 464.70 examples/s]" + "value": "\u2007100000/100000\u2007[04:18<00:00,\u2007464.70\u2007examples/s]" } }, "4d4f8726c88d40bcabdb664a663cbcc8": { @@ -3670,9 +3669,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c88063243cb0423db943d5c14b53b799", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_af75f290b4e1431681085b82b894129c", - "value": " 3.08G/3.08G [00:34<00:00, 244MB/s]" + "value": "\u20073.08G/3.08G\u2007[00:34<00:00,\u2007244MB/s]" } }, "5471cb70b9fc476e924a79bae6fbe973": { @@ -3871,9 +3870,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_63a2c8a60c494acd87b00b4cfdff5c68", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_adbddacbb5414d24a37c1a4a0f036547", - "value": "tokenizer.json: 100%" + "value": "value": "tokenizer.json:\u2007100%" } }, "564ec00e94eb477785335399eb492556": { @@ -3892,9 +3891,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1aba00c5303e49d88c4bc8987c8e2b03", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_bf150441841847489593ab9cffc385ac", - "value": "chat_template.jinja: " + "value": "value": "chat_template.jinja:\u2007" } }, "573ad1a363b746699575ad8d77592312": { @@ -4452,9 +4451,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fc3356580e6744ca83b8649368480b3e", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_5809cbdec1844cbebc0db710185d8072", - "value": "Unsloth: Standardizing formats (num_proc=2): 100%" + "value": "Unsloth:\u2007Standardizing\u2007formats\u2007(num_proc=2):\u2007100%" } }, "63a2c8a60c494acd87b00b4cfdff5c68": { @@ -4593,9 +4592,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_f72b1f631cc6458dbbdbc4fc3642133f", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_c0a0d7c3181442c69fd330ed1452dd84", - "value": "merges.txt: " + "value": "merges.txt:\u2007" } }, "67c3227617ea4f72a5a3a9e822bb8abb": { @@ -5146,9 +5145,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a1cba3ed46c34f2ebfaf2b1dbb72d0d3", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_573ad1a363b746699575ad8d77592312", - "value": "vocab.json: " + "value": "vocab.json:\u2007" } }, "803af90c303241edbbca3338dd4f043d": { @@ -5234,9 +5233,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_613103c611fd412c815a37de1007dd3d", - "placeholder": "​", + "placeholder": "​\u200b", "style": "IPY_MODEL_78a08d6e960f43ad9b74f46c125b5aea", - "value": "data/train-00000-of-00001.parquet: 100%" + "value": "value": "data/train-00000-of-00001.parquet:\u2007100%" } }, "854ab752c85449ad80b68ad7677df198": { @@ -5307,9 +5306,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_598cdec8e8dd49a0bec21b74394ad47d", - "placeholder": "​", + "placeholder": "​\u200b", "style": "IPY_MODEL_da70610720a14a1ca0e689850a081809", - "value": "tokenizer_config.json: " + "value": "tokenizer_config.json:\u2007" } }, "8e0aa0bf9fdf456bae3ff6dae9825eac": { @@ -5472,9 +5471,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6226ba2708f54d3682c2257422299b61", - "placeholder": "​", + "placeholder": "​\u200b", "style": "IPY_MODEL_5b790390fed14b339b58be310eaf5b7b", - "value": " 614/614 [00:00<00:00, 13.3kB/s]" + "value": "\u2007614/614\u2007[00:00<00:00,\u200713.3kB/s]" } }, "94561bbb46cc4636ab21428a8a47f107": { @@ -5508,9 +5507,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_40b609ebe3c2437a9ec895e48de6d1ca", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_6e7a0eb4f9e340bd9883bc975c6a489c", - "value": " 707/707 [00:00<00:00, 10.8kB/s]" + "value": "\u2007707/707\u2007[00:00<00:00,\u200710.8kB/s]" } }, "984b8bee3e6e493e8bebaa177e9a8aa3": { @@ -5597,9 +5596,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ba99aaf582914c83b18910ac713a9b6f", - "placeholder": "​", + "placeholder": "​\u200b", "style": "IPY_MODEL_94561bbb46cc4636ab21428a8a47f107", - "value": " 100000/100000 [00:32<00:00, 2848.96 examples/s]" + "value": "\u2007100000/100000\u2007[00:32<00:00,\u20072848.96\u2007examples/s]" } }, "9a89009072e647b2b760688428c76e06": { @@ -5746,9 +5745,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fa846c3ef9124e6090530fd0ad72851e", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_3ab7781489bc4bbe9b072fd49f7c12f8", - "value": "Map (num_proc=6): 100%" + "value": "Map\u2007(num_proc=6):\u2007100%" } }, "9efb3a9c377c4ce3bf13038cd8cb9c63": { @@ -5999,9 +5998,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_61c268a7d13a44afa56bedade90e4f34", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_272afad7361c4eeab631c7bd41e6fa14", - "value": " 4.97G/4.97G [00:57<00:00, 263MB/s]" + "value": "\u20074.97G/4.97G\u2007[00:57<00:00,\u2007263MB/s]" } }, "a893ae8f4f954f78926d91a2a2b019de": { @@ -6088,9 +6087,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0e31ade8b833471a81309de8c821bce9", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_88b7cea7ba0e42b7866bbfa4dc2df056", - "value": " 100000/100000 [00:01<00:00, 97725.37 examples/s]" + "value": "\u2007100000/100000\u2007[00:01<00:00,\u200797725.37\u2007examples/s]" } }, "a95fca80392b4af3aae5d189059dace6": { @@ -6109,9 +6108,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4641a1c322e94997a1d4e99ba4d29db4", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_e2e1b4e39b7c44caaf698082cef96f20", - "value": "Unsloth: Tokenizing ["text"] (num_proc=6): 100%" + "value": "Unsloth:\u2007Tokenizing\u2007["text"]\u2007(num_proc=6):\u2007100%" } }, "a98adcf1d23547f599e01ab83bec8b48": { @@ -6250,9 +6249,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c993d2d141094cf0b7151a8dbffd9a45", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_3d9701a29b684fb0ac32535accf25ccb", - "value": " 2.78M/? [00:00<00:00, 12.3MB/s]" + "value": "\u20072.78M/?\u2007[00:00<00:00,\u200712.3MB/s]" } }, "ac2ceaa6d5fe41169557e673ccc83ec5": { @@ -6375,9 +6374,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_cc8af4f890a54538ba6aae19866dc7c0", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_803af90c303241edbbca3338dd4f043d", - "value": " 11.4M/11.4M [00:01<00:00, 9.26MB/s]" + "value": "\u200711.4M/11.4M\u2007[00:01<00:00,\u20079.26MB/s]" } }, "affc3e6798984ab7bee3b100e9bf07f3": { @@ -6396,9 +6395,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_79ac14919f7c47de914f634481c7c47a", - "placeholder": "​", + "placeholder": "​\u200b", "style": "IPY_MODEL_611831076b59469fbcfcacdec4ac4bea", - "value": " 32.9k/? [00:00<00:00, 3.04MB/s]" + "value": "\u200732.9k/?\u2007[00:00<00:00,\u20073.04MB/s]" } }, "b021cc01ce3748d48aa7fc32843c51f9": { @@ -6958,9 +6957,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1ea6c960f5d3449eb445ba26b226ab80", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_b7a3d157a73644c899f1e4d60678d9c3", - "value": "special_tokens_map.json: 100%" + "value": "special_tokens_map.json:\u2007100%" } }, "c5be281628bd49f0b184cd899143c2c4": { @@ -7426,9 +7425,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_592370835995436e8ce573f9f121b410", - "placeholder": "​", + "placeholder": "​\u200b", "style": "IPY_MODEL_24d4ad43783e4715b89a1bbd0a84b90c", - "value": "Map: 100%" + "value": "Map:\u2007100%" } }, "d38043a7e0ce4eeabd900a7ad01fb733": { @@ -7469,9 +7468,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a6c5b152ce8a4845aab8f4bd42bbc081", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_2a86e8d4b9d64fc990de342af2359327", - "value": "Loading checkpoint shards: 100%" + "value": "Loading\u2007checkpoint\u2007shards:\u2007100%" } }, "d5ac21ecf53146b9a91e77c1365a8087": { @@ -7649,9 +7648,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a8f346bd2311473bb41f9c832d38fde3", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_27080c04e0f44a21986825766793e12a", - "value": "generation_config.json: 100%" + "value": "generation_config.json:\u2007100%" } }, "dfe5475d5906460e9a771e0eeefdcd46": { @@ -7670,9 +7669,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_7648f19accea41f5abec2f3d8296ac19", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_a34d8312550e4c89a37dafb88a906dcd", - "value": " 117M/117M [00:01<00:00, 118MB/s]" + "value": "\u2007117M/117M\u2007[00:01<00:00,\u2007118MB/s]" } }, "e01669badd9043e998951c212c8ff174": { @@ -7880,9 +7879,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6eecd84361254cf6973f16e70265707b", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_0fa3c3feb88042afbd921a51c91353ee", - "value": "added_tokens.json: 100%" + "value": "added_tokens.json:\u2007100%" } }, "e86016cd0c124ca78fe3119e59c94de8": { @@ -7923,9 +7922,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_3ff4a0687fdf44fa95afe9ec0ab781ea", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_a461d5efcb48465eb6c1b39a7d67fa07", - "value": "README.md: 100%" + "value": "README.md:\u2007100%" } }, "ea888f9b76b448208a11d6eee0c1eb81": { @@ -8185,9 +8184,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c5be281628bd49f0b184cd899143c2c4", - "placeholder": "​", + "placeholder": "​\u200b", "style": "IPY_MODEL_4d4f8726c88d40bcabdb664a663cbcc8", - "value": "model-00001-of-00002.safetensors: 100%" + "value": "model-00001-of-00002.safetensors:\u2007100%" } }, "f5279e94f100416e858be5122226ef53": { @@ -8280,9 +8279,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_577decf0054b409a95ff0a0bf23b73ed", - "placeholder": "​", + "placeholder": "\u200b​", "style": "IPY_MODEL_cce32a3c7860476690a0e5762f22b1ab", - "value": " 1.67M/? [00:00<00:00, 7.39MB/s]" + "value": "\u20071.67M/?\u2007[00:00<00:00,\u20077.39MB/s]" } }, "f81c5ed93c684019b289c7a01ec0f177": { From d8dffa6a9bd5b0d7ab2b5dd634977ef4a682b368 Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 10:04:14 +0200 Subject: [PATCH 03/12] Update Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb | 31 +++++++------------------ 1 file changed, 8 insertions(+), 23 deletions(-) diff --git a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb index 64956a9b..e89e5a89 100644 --- a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb @@ -51,22 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": [ - "%%capture\n", - "import os, re\n", - "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", - " !pip install unsloth\n", - "else:\n", - " # Do this only in Colab notebooks! Otherwise use pip install unsloth\n", - " import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n", - " xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n", - " !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n", - " !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n", - " !pip install --no-deps unsloth\n", - "!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n", - "!pip install transformers==4.55.4\n", - "!pip install --no-deps trl==0.22.2" - ] + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown", @@ -1654,7 +1639,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + \u2b50\ufe0f Star us on Github ⭐️\n", + " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", "\n", " This notebook and all Unsloth notebooks are licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme)\n", "\n" @@ -2673,7 +2658,7 @@ "layout": "IPY_MODEL_1c977d30200d437abcc3bab5459ac3a0", "placeholder": "\u200b", "style": "IPY_MODEL_edc64ecdba034ac5afc978bf8c844016", - "\u20072/2\u2007[00:31<00:00,\u200714.00s/it]" + "value": "\u20072/2\u2007[00:31<00:00,\u200714.00s/it]" } }, "23182fb1edd248259238b5eae55811e2": { @@ -3191,7 +3176,7 @@ "layout": "IPY_MODEL_b7f2a83d93994f83b4163627a6619c72", "placeholder": "\u200b", "style": "IPY_MODEL_47586e1fc8cb454e84d447d3852fb8fc", - "\u20072/2\u2007[00:44<00:00,\u200721.28s/it]" + "value": "\u20072/2\u2007[00:44<00:00,\u200721.28s/it]" } }, "40b609ebe3c2437a9ec895e48de6d1ca": { @@ -3441,7 +3426,7 @@ "layout": "IPY_MODEL_54bb3cb7e89843e881366b277107f10e", "placeholder": "\u200b", "style": "IPY_MODEL_b48febda135143208ab19ddbc54db83f", - "value": "value": "\u20074.04k/?\u2007[00:00<00:00,\u2007405kB/s]" + "value": "\u20074.04k/?\u2007[00:00<00:00,\u2007405kB/s]" } }, "47586e1fc8cb454e84d447d3852fb8fc": { @@ -3872,7 +3857,7 @@ "layout": "IPY_MODEL_63a2c8a60c494acd87b00b4cfdff5c68", "placeholder": "\u200b​", "style": "IPY_MODEL_adbddacbb5414d24a37c1a4a0f036547", - "value": "value": "tokenizer.json:\u2007100%" + "value": "tokenizer.json:\u2007100%" } }, "564ec00e94eb477785335399eb492556": { @@ -3893,7 +3878,7 @@ "layout": "IPY_MODEL_1aba00c5303e49d88c4bc8987c8e2b03", "placeholder": "\u200b​", "style": "IPY_MODEL_bf150441841847489593ab9cffc385ac", - "value": "value": "chat_template.jinja:\u2007" + "value": "chat_template.jinja:\u2007" } }, "573ad1a363b746699575ad8d77592312": { @@ -5235,7 +5220,7 @@ "layout": "IPY_MODEL_613103c611fd412c815a37de1007dd3d", "placeholder": "​\u200b", "style": "IPY_MODEL_78a08d6e960f43ad9b74f46c125b5aea", - "value": "value": "data/train-00000-of-00001.parquet:\u2007100%" + "value": "data/train-00000-of-00001.parquet:\u2007100%" } }, "854ab752c85449ad80b68ad7677df198": { From e243d68b9e18890cddfa9237d726bcbb3b87bee5 Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 10:25:33 +0200 Subject: [PATCH 04/12] Update Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Qwen3_(4B)_Instruct-QAT.ipynb | 194 ++++++++++++++----------------- 1 file changed, 89 insertions(+), 105 deletions(-) diff --git a/nb/Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Qwen3_(4B)_Instruct-QAT.ipynb index 60d45f14..81ab31dc 100644 --- a/nb/Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Qwen3_(4B)_Instruct-QAT.ipynb @@ -8,7 +8,7 @@ "
\n", "\n", "\n", - " Join Discord if you need help + ⭐ Star us on Github ⭐\n", + " Join Discord if you need help + \u2b50 Star us on Github \u2b50\n", "
\n", "\n", "To install Unsloth your local device, follow [our guide](https://docs.unsloth.ai/get-started/install-and-update). This notebook is licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme).\n", @@ -51,22 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": [ - "%%capture\n", - "import os, re\n", - "if \"COLAB_\" not in \"\".join(os.environ.keys()):\n", - " !pip install unsloth\n", - "else:\n", - " # Do this only in Colab notebooks! Otherwise use pip install unsloth\n", - " import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n", - " xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n", - " !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n", - " !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n", - " !pip install --no-deps unsloth\n", - "!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n", - "!pip install transformers==4.55.4\n", - "!pip install --no-deps trl==0.22.2" - ] + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown", @@ -227,7 +212,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "πŸ¦₯ Unsloth: Will patch your computer to enable 2x faster free finetuning.\n" + "\ud83e\udda5 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n" ] }, { @@ -242,7 +227,7 @@ "name": "stdout", "output_type": "stream", "text": [ - "πŸ¦₯ Unsloth Zoo will now patch everything to make training faster!\n", + "\ud83e\udda5 Unsloth Zoo will now patch everything to make training faster!\n", "==((====))== Unsloth 2025.10.8: Fast Qwen3 patching. Transformers: 4.55.4.\n", " \\\\ /| Tesla T4. Num GPUs = 1. Max memory: 14.741 GB. Platform: Linux.\n", "O^O/ \\_/ \\ Torch: 2.8.0+cu126. CUDA: 7.5. CUDA Toolkit: 12.6. Triton: 3.4.0\n", @@ -1534,7 +1519,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "Unsloth: Copying 2 files from cache to `model`: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [03:20<00:00, 100.23s/it]\n" + "Unsloth: Copying 2 files from cache to `model`: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [03:20<00:00, 100.23s/it]\n" ] }, { @@ -1551,8 +1536,8 @@ "name": "stderr", "output_type": "stream", "text": [ - "Unsloth: Preparing safetensor model files: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [00:00<00:00, 15709.00it/s]\n", - "Unsloth: Merging weights into 16bit: 100%|β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆ| 2/2 [04:31<00:00, 135.70s/it]\n" + "Unsloth: Preparing safetensor model files: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [00:00<00:00, 15709.00it/s]\n", + "Unsloth: Merging weights into 16bit: 100%|\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588\u2588| 2/2 [04:31<00:00, 135.70s/it]\n" ] }, { @@ -1654,7 +1639,7 @@ " \n", " \n", "\n", - " Join Discord if you need help + ⭐️ Star us on Github ⭐️\n", + " Join Discord if you need help + \u2b50\ufe0f Star us on Github \u2b50\ufe0f\n", "\n", " This notebook and all Unsloth notebooks are licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme)\n", "\n" @@ -1669,8 +1654,7 @@ "provenance": [] }, "kernelspec": { - "display_name": "Python 3 (ipykernel)", - "language": "python", + "display_name": "Python 3", "name": "python3" }, "language_info": { @@ -1703,9 +1687,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_5471cb70b9fc476e924a79bae6fbe973", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_10ccee1030c94c05a299e4dbb646f26c", - "value": " 100000/100000 [00:52<00:00, 1750.73 examples/s]" + "value": "\u2007100000/100000\u2007[00:52<00:00,\u20071750.73\u2007examples/s]" } }, "0297fab8f99f4de6a905c5e16ecf6d45": { @@ -1798,9 +1782,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c168576851f240a9a6f0c9de5045ea76", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_6f4d50b5f5db4d27a3219b6a83560542", - "value": " 9.65k/? [00:00<00:00, 164kB/s]" + "value": "\u20079.65k/?\u2007[00:00<00:00,\u2007164kB/s]" } }, "0658d6cd875e4ddfbcaec38acfc6ab8e": { @@ -2210,9 +2194,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_d027fe1c1c874b00b35d1aa728e2f64b", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_c050f260067d4a8cb0a48b8789853e42", - "value": "model.safetensors.index.json: " + "value": "model.safetensors.index.json:\u2007" } }, "15493837f6ed4859a856749581e4bc94": { @@ -2299,9 +2283,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_eed8825725c8493aa839ee1f662981a8", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_873775127eff456e8418e310c10f68c6", - "value": "Generating train split: 100%" + "value": "Generating\u2007train\u2007split:\u2007100%" } }, "1aba00c5303e49d88c4bc8987c8e2b03": { @@ -2672,9 +2656,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1c977d30200d437abcc3bab5459ac3a0", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_edc64ecdba034ac5afc978bf8c844016", - "value": " 2/2 [00:31<00:00, 14.00s/it]" + "value": "\u20072/2\u2007[00:31<00:00,\u200714.00s/it]" } }, "23182fb1edd248259238b5eae55811e2": { @@ -2745,9 +2729,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c2986f10f6a64a2fa7bd4fdb70e08e04", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_9efb3a9c377c4ce3bf13038cd8cb9c63", - "value": " 100000/100000 [00:04<00:00, 17940.57 examples/s]" + "value": "\u2007100000/100000\u2007[00:04<00:00,\u200717940.57\u2007examples/s]" } }, "24d4ad43783e4715b89a1bbd0a84b90c": { @@ -2827,9 +2811,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_74499d604e454b009f07f9900f010e4e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_3312b27851c64b50a3e52013b4285c13", - "value": " 982/982 [00:00<00:00, 114kB/s]" + "value": "\u2007982/982\u2007[00:00<00:00,\u2007114kB/s]" } }, "2a86e8d4b9d64fc990de342af2359327": { @@ -2952,9 +2936,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c8de0d52dd214fd283bf3d2ff9c50bd7", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_a0f7475651e84e02afd306a2f6d8cb51", - "value": "Loading checkpoint shards: 100%" + "value": "Loading\u2007checkpoint\u2007shards:\u2007100%" } }, "36986d45f3f342f494287555aab6d79c": { @@ -3117,9 +3101,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_36986d45f3f342f494287555aab6d79c", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_7ad06c11c12e4d12bb8567f46e005940", - "value": " 237/237 [00:00<00:00, 6.90kB/s]" + "value": "\u2007237/237\u2007[00:00<00:00,\u20076.90kB/s]" } }, "3ff4a0687fdf44fa95afe9ec0ab781ea": { @@ -3190,9 +3174,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_b7f2a83d93994f83b4163627a6619c72", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_47586e1fc8cb454e84d447d3852fb8fc", - "value": " 2/2 [00:44<00:00, 21.28s/it]" + "value": "\u20072/2\u2007[00:44<00:00,\u200721.28s/it]" } }, "40b609ebe3c2437a9ec895e48de6d1ca": { @@ -3367,9 +3351,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4b78a2ab29ba40bb98e18bf3beb1181e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_5d6d1d2d76dd45d4a6ec9d1cc067968c", - "value": "model-00002-of-00002.safetensors: 100%" + "value": "model-00002-of-00002.safetensors:\u2007100%" } }, "4641a1c322e94997a1d4e99ba4d29db4": { @@ -3440,9 +3424,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_54bb3cb7e89843e881366b277107f10e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_b48febda135143208ab19ddbc54db83f", - "value": " 4.04k/? [00:00<00:00, 405kB/s]" + "value": "\u20074.04k/?\u2007[00:00<00:00,\u2007405kB/s]" } }, "47586e1fc8cb454e84d447d3852fb8fc": { @@ -3566,9 +3550,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ea93bb20bbc44092ab0b1394cbd28d4e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_06d793c0fe3f46e29e279674de222b8f", - "value": " 100000/100000 [04:18<00:00, 464.70 examples/s]" + "value": "\u2007100000/100000\u2007[04:18<00:00,\u2007464.70\u2007examples/s]" } }, "4d4f8726c88d40bcabdb664a663cbcc8": { @@ -3670,9 +3654,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c88063243cb0423db943d5c14b53b799", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_af75f290b4e1431681085b82b894129c", - "value": " 3.08G/3.08G [00:34<00:00, 244MB/s]" + "value": "\u20073.08G/3.08G\u2007[00:34<00:00,\u2007244MB/s]" } }, "5471cb70b9fc476e924a79bae6fbe973": { @@ -3871,9 +3855,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_63a2c8a60c494acd87b00b4cfdff5c68", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_adbddacbb5414d24a37c1a4a0f036547", - "value": "tokenizer.json: 100%" + "value": "tokenizer.json:\u2007100%" } }, "564ec00e94eb477785335399eb492556": { @@ -3892,9 +3876,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1aba00c5303e49d88c4bc8987c8e2b03", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_bf150441841847489593ab9cffc385ac", - "value": "chat_template.jinja: " + "value": "chat_template.jinja:\u2007" } }, "573ad1a363b746699575ad8d77592312": { @@ -4452,9 +4436,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fc3356580e6744ca83b8649368480b3e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_5809cbdec1844cbebc0db710185d8072", - "value": "Unsloth: Standardizing formats (num_proc=2): 100%" + "value": "Unsloth:\u2007Standardizing\u2007formats\u2007(num_proc=2):\u2007100%" } }, "63a2c8a60c494acd87b00b4cfdff5c68": { @@ -4593,9 +4577,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_f72b1f631cc6458dbbdbc4fc3642133f", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_c0a0d7c3181442c69fd330ed1452dd84", - "value": "merges.txt: " + "value": "merges.txt:\u2007" } }, "67c3227617ea4f72a5a3a9e822bb8abb": { @@ -5146,9 +5130,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a1cba3ed46c34f2ebfaf2b1dbb72d0d3", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_573ad1a363b746699575ad8d77592312", - "value": "vocab.json: " + "value": "vocab.json:\u2007" } }, "803af90c303241edbbca3338dd4f043d": { @@ -5234,9 +5218,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_613103c611fd412c815a37de1007dd3d", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_78a08d6e960f43ad9b74f46c125b5aea", - "value": "data/train-00000-of-00001.parquet: 100%" + "value": "data/train-00000-of-00001.parquet:\u2007100%" } }, "854ab752c85449ad80b68ad7677df198": { @@ -5307,9 +5291,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_598cdec8e8dd49a0bec21b74394ad47d", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_da70610720a14a1ca0e689850a081809", - "value": "tokenizer_config.json: " + "value": "tokenizer_config.json:\u2007" } }, "8e0aa0bf9fdf456bae3ff6dae9825eac": { @@ -5472,9 +5456,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6226ba2708f54d3682c2257422299b61", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_5b790390fed14b339b58be310eaf5b7b", - "value": " 614/614 [00:00<00:00, 13.3kB/s]" + "value": "\u2007614/614\u2007[00:00<00:00,\u200713.3kB/s]" } }, "94561bbb46cc4636ab21428a8a47f107": { @@ -5508,9 +5492,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_40b609ebe3c2437a9ec895e48de6d1ca", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_6e7a0eb4f9e340bd9883bc975c6a489c", - "value": " 707/707 [00:00<00:00, 10.8kB/s]" + "value": "\u2007707/707\u2007[00:00<00:00,\u200710.8kB/s]" } }, "984b8bee3e6e493e8bebaa177e9a8aa3": { @@ -5597,9 +5581,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_ba99aaf582914c83b18910ac713a9b6f", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_94561bbb46cc4636ab21428a8a47f107", - "value": " 100000/100000 [00:32<00:00, 2848.96 examples/s]" + "value": "\u2007100000/100000\u2007[00:32<00:00,\u20072848.96\u2007examples/s]" } }, "9a89009072e647b2b760688428c76e06": { @@ -5746,9 +5730,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_fa846c3ef9124e6090530fd0ad72851e", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_3ab7781489bc4bbe9b072fd49f7c12f8", - "value": "Map (num_proc=6): 100%" + "value": "Map\u2007(num_proc=6):\u2007100%" } }, "9efb3a9c377c4ce3bf13038cd8cb9c63": { @@ -5999,9 +5983,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_61c268a7d13a44afa56bedade90e4f34", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_272afad7361c4eeab631c7bd41e6fa14", - "value": " 4.97G/4.97G [00:57<00:00, 263MB/s]" + "value": "\u20074.97G/4.97G\u2007[00:57<00:00,\u2007263MB/s]" } }, "a893ae8f4f954f78926d91a2a2b019de": { @@ -6088,9 +6072,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_0e31ade8b833471a81309de8c821bce9", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_88b7cea7ba0e42b7866bbfa4dc2df056", - "value": " 100000/100000 [00:01<00:00, 97725.37 examples/s]" + "value": "\u2007100000/100000\u2007[00:01<00:00,\u200797725.37\u2007examples/s]" } }, "a95fca80392b4af3aae5d189059dace6": { @@ -6109,9 +6093,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_4641a1c322e94997a1d4e99ba4d29db4", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_e2e1b4e39b7c44caaf698082cef96f20", - "value": "Unsloth: Tokenizing ["text"] (num_proc=6): 100%" + "value": "Unsloth:\u2007Tokenizing\u2007["text"]\u2007(num_proc=6):\u2007100%" } }, "a98adcf1d23547f599e01ab83bec8b48": { @@ -6250,9 +6234,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c993d2d141094cf0b7151a8dbffd9a45", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_3d9701a29b684fb0ac32535accf25ccb", - "value": " 2.78M/? [00:00<00:00, 12.3MB/s]" + "value": "\u20072.78M/?\u2007[00:00<00:00,\u200712.3MB/s]" } }, "ac2ceaa6d5fe41169557e673ccc83ec5": { @@ -6375,9 +6359,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_cc8af4f890a54538ba6aae19866dc7c0", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_803af90c303241edbbca3338dd4f043d", - "value": " 11.4M/11.4M [00:01<00:00, 9.26MB/s]" + "value": "\u200711.4M/11.4M\u2007[00:01<00:00,\u20079.26MB/s]" } }, "affc3e6798984ab7bee3b100e9bf07f3": { @@ -6396,9 +6380,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_79ac14919f7c47de914f634481c7c47a", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_611831076b59469fbcfcacdec4ac4bea", - "value": " 32.9k/? [00:00<00:00, 3.04MB/s]" + "value": "\u200732.9k/?\u2007[00:00<00:00,\u20073.04MB/s]" } }, "b021cc01ce3748d48aa7fc32843c51f9": { @@ -6958,9 +6942,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_1ea6c960f5d3449eb445ba26b226ab80", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_b7a3d157a73644c899f1e4d60678d9c3", - "value": "special_tokens_map.json: 100%" + "value": "special_tokens_map.json:\u2007100%" } }, "c5be281628bd49f0b184cd899143c2c4": { @@ -7426,9 +7410,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_592370835995436e8ce573f9f121b410", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_24d4ad43783e4715b89a1bbd0a84b90c", - "value": "Map: 100%" + "value": "Map:\u2007100%" } }, "d38043a7e0ce4eeabd900a7ad01fb733": { @@ -7469,9 +7453,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a6c5b152ce8a4845aab8f4bd42bbc081", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_2a86e8d4b9d64fc990de342af2359327", - "value": "Loading checkpoint shards: 100%" + "value": "Loading\u2007checkpoint\u2007shards:\u2007100%" } }, "d5ac21ecf53146b9a91e77c1365a8087": { @@ -7649,9 +7633,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_a8f346bd2311473bb41f9c832d38fde3", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_27080c04e0f44a21986825766793e12a", - "value": "generation_config.json: 100%" + "value": "generation_config.json:\u2007100%" } }, "dfe5475d5906460e9a771e0eeefdcd46": { @@ -7670,9 +7654,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_7648f19accea41f5abec2f3d8296ac19", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_a34d8312550e4c89a37dafb88a906dcd", - "value": " 117M/117M [00:01<00:00, 118MB/s]" + "value": "\u2007117M/117M\u2007[00:01<00:00,\u2007118MB/s]" } }, "e01669badd9043e998951c212c8ff174": { @@ -7880,9 +7864,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_6eecd84361254cf6973f16e70265707b", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_0fa3c3feb88042afbd921a51c91353ee", - "value": "added_tokens.json: 100%" + "value": "added_tokens.json:\u2007100%" } }, "e86016cd0c124ca78fe3119e59c94de8": { @@ -7923,9 +7907,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_3ff4a0687fdf44fa95afe9ec0ab781ea", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_a461d5efcb48465eb6c1b39a7d67fa07", - "value": "README.md: 100%" + "value": "README.md:\u2007100%" } }, "ea888f9b76b448208a11d6eee0c1eb81": { @@ -8185,9 +8169,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_c5be281628bd49f0b184cd899143c2c4", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_4d4f8726c88d40bcabdb664a663cbcc8", - "value": "model-00001-of-00002.safetensors: 100%" + "value": "model-00001-of-00002.safetensors:\u2007100%" } }, "f5279e94f100416e858be5122226ef53": { @@ -8280,9 +8264,9 @@ "description": "", "description_tooltip": null, "layout": "IPY_MODEL_577decf0054b409a95ff0a0bf23b73ed", - "placeholder": "​", + "placeholder": "\u200b", "style": "IPY_MODEL_cce32a3c7860476690a0e5762f22b1ab", - "value": " 1.67M/? [00:00<00:00, 7.39MB/s]" + "value": "\u20071.67M/?\u2007[00:00<00:00,\u20077.39MB/s]" } }, "f81c5ed93c684019b289c7a01ec0f177": { From de2401de6109668f293e4d139725cba11e9ee684 Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 10:27:14 +0200 Subject: [PATCH 05/12] Update Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Qwen3_(4B)_Instruct-QAT.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/nb/Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Qwen3_(4B)_Instruct-QAT.ipynb index 81ab31dc..c7d18488 100644 --- a/nb/Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Qwen3_(4B)_Instruct-QAT.ipynb @@ -51,7 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n !pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n !pip install transformers==4.55.4\n !pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown", From 9315394b18d2994c0bf5b2836f66f8231e38e19b Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 10:29:15 +0200 Subject: [PATCH 06/12] Update Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb From 3c0feed15a34fe8a0be669313e626146fa399158 Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 11:01:57 +0200 Subject: [PATCH 07/12] Update Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb index e89e5a89..fb74364f 100644 --- a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb @@ -51,7 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown", From a41409ddd2f52eaf32cbc9d7e547664db45cb8b7 Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 11:03:11 +0200 Subject: [PATCH 08/12] Update Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb index fb74364f..3a42b0ae 100644 --- a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb @@ -51,7 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown", From de6f1185d94ca8cf8d41b1361479f9bfeeec50ec Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 11:04:25 +0200 Subject: [PATCH 09/12] Update Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb index 3a42b0ae..88f71900 100644 --- a/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Kaggle-Qwen3_(4B)_Instruct-QAT.ipynb @@ -51,7 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown", From a3f3722a560489af27227c234b17f11b53479c5e Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 11:05:36 +0200 Subject: [PATCH 10/12] Update Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Qwen3_(4B)_Instruct-QAT.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/nb/Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Qwen3_(4B)_Instruct-QAT.ipynb index c7d18488..e62377aa 100644 --- a/nb/Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Qwen3_(4B)_Instruct-QAT.ipynb @@ -51,7 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n !pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n !pip install transformers==4.55.4\n !pip install --no-deps trl==0.22.2" + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n !pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n !pip install transformers==4.55.4\n !pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown", From dcce0bd68c5bfe59141774ac3ec8f36c0ff7a8da Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 11:06:03 +0200 Subject: [PATCH 11/12] Update Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Qwen3_(4B)_Instruct-QAT.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/nb/Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Qwen3_(4B)_Instruct-QAT.ipynb index e62377aa..a9b992d2 100644 --- a/nb/Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Qwen3_(4B)_Instruct-QAT.ipynb @@ -51,7 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n !pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n !pip install transformers==4.55.4\n !pip install --no-deps trl==0.22.2" + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n !pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n !pip install transformers==4.55.4\n !pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown", From 7d7bb1e720c2ed651d1832c643816df8cbda0681 Mon Sep 17 00:00:00 2001 From: Roland Tannous <115670425+rolandtannous@users.noreply.github.com> Date: Tue, 18 Nov 2025 11:07:03 +0200 Subject: [PATCH 12/12] Update Qwen3_(4B)_Instruct-QAT.ipynb --- nb/Qwen3_(4B)_Instruct-QAT.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/nb/Qwen3_(4B)_Instruct-QAT.ipynb b/nb/Qwen3_(4B)_Instruct-QAT.ipynb index a9b992d2..c7071be2 100644 --- a/nb/Qwen3_(4B)_Instruct-QAT.ipynb +++ b/nb/Qwen3_(4B)_Instruct-QAT.ipynb @@ -51,7 +51,7 @@ "execution_count": null, "metadata": {}, "outputs": [], - "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n !pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n !pip install transformers==4.55.4\n !pip install --no-deps trl==0.22.2" + "source": "%%capture\nimport os, re\nif \"COLAB_\" not in \"\".join(os.environ.keys()):\n !pip install unsloth\nelse:\n # Do this only in Colab notebooks! Otherwise use pip install unsloth\n import torch; v = re.match(r\"[0-9\\.]{3,}\", str(torch.__version__)).group(0)\n xformers = \"xformers==\" + (\"0.0.32.post2\" if v == \"2.8.0\" else \"0.0.29.post3\")\n !pip install --no-deps bitsandbytes accelerate {xformers} peft trl triton cut_cross_entropy unsloth_zoo\n !pip install sentencepiece protobuf \"datasets>=3.4.1,<4.0.0\" \"huggingface_hub>=0.34.0\" hf_transfer\n !pip install --no-deps unsloth\n!pip install torchao==0.14.0 fbgemm-gpu-genai==1.3.0\n!pip install transformers==4.55.4\n!pip install --no-deps trl==0.22.2" }, { "cell_type": "markdown",