From f32b361c8e3febb79cc25330e85d564eccc63872 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Monk=20=28looking=20for=20PhD=20Fall=E2=80=9924=29?=
 <aadityaura@gmail.com>
Date: Thu, 18 Apr 2024 22:43:33 +0530
Subject: [PATCH 1/2] Create qlora.yml

---
 examples/llama-3/qlora.yml | 66 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 66 insertions(+)
 create mode 100644 examples/llama-3/qlora.yml

diff --git a/examples/llama-3/qlora.yml b/examples/llama-3/qlora.yml
new file mode 100644
index 000000000..3c81cd7b7
--- /dev/null
+++ b/examples/llama-3/qlora.yml
@@ -0,0 +1,66 @@
+base_model: meta-llama/Meta-Llama-3-8B
+model_type: AutoModelForCausalLM
+tokenizer_type: AutoTokenizer
+
+load_in_8bit: false
+load_in_4bit: true
+strict: false
+
+datasets:
+  - path: aaditya/alpaca_subset_1
+    type: alpaca
+dataset_prepared_path:
+val_set_size: 0.05
+output_dir: ./qlora-out
+
+adapter: qlora
+lora_model_dir:
+
+sequence_len: 4096
+sample_packing: true
+pad_to_sequence_len: true
+
+lora_r: 32
+lora_alpha: 16
+lora_dropout: 0.05
+lora_target_modules:
+lora_target_linear: true
+lora_fan_in_fan_out:
+
+wandb_project:
+wandb_entity:
+wandb_watch:
+wandb_name:
+wandb_log_model:
+
+gradient_accumulation_steps: 4
+micro_batch_size: 2
+num_epochs: 4
+optimizer: paged_adamw_32bit
+lr_scheduler: cosine
+learning_rate: 0.0002
+
+train_on_inputs: false
+group_by_length: false
+bf16: auto
+fp16:
+tf32: false
+
+gradient_checkpointing: true
+early_stopping_patience:
+resume_from_checkpoint:
+local_rank:
+logging_steps: 1
+xformers_attention:
+flash_attention: true
+
+warmup_steps: 10
+evals_per_epoch: 4
+eval_table_size:
+saves_per_epoch: 1
+debug:
+deepspeed:
+weight_decay: 0.0
+fsdp:
+fsdp_config:
+special_tokens:

From 031d83b30d690a0e6801a557f7bb6675ff0e1627 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Monk=20=28looking=20for=20PhD=20Fall=E2=80=9924=29?=
 <aadityaura@gmail.com>
Date: Thu, 18 Apr 2024 23:02:57 +0530
Subject: [PATCH 2/2] Update qlora.yml

---
 examples/llama-3/qlora.yml | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/examples/llama-3/qlora.yml b/examples/llama-3/qlora.yml
index 3c81cd7b7..9cedee8ee 100644
--- a/examples/llama-3/qlora.yml
+++ b/examples/llama-3/qlora.yml
@@ -10,7 +10,7 @@ datasets:
   - path: aaditya/alpaca_subset_1
     type: alpaca
 dataset_prepared_path:
-val_set_size: 0.05
+val_set_size: 0
 output_dir: ./qlora-out
 
 adapter: qlora
@@ -64,3 +64,4 @@ weight_decay: 0.0
 fsdp:
 fsdp_config:
 special_tokens:
+  pad_token: "<|end_of_text|>"