-
Notifications
You must be signed in to change notification settings - Fork 301
Mirror all weights on HF from Kaggle #1959
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Merged
divyashreepathihalli
merged 7 commits into
keras-team:master
from
divyashreepathihalli:mirror_hf
Oct 30, 2024
Merged
Changes from all commits
Commits
Show all changes
7 commits
Select commit
Hold shift + click to select a range
aaec125
Mirror all weights on HF
divyashreepathihalli 92e8ca2
save latest version of preset list
divyashreepathihalli 932a899
clean up
divyashreepathihalli 53f46ae
add try block
divyashreepathihalli f064877
improve print and error message
divyashreepathihalli 549ab73
update the final json file
divyashreepathihalli f6aae4b
update presets
divyashreepathihalli File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,149 @@ | ||
| [ | ||
| "kaggle://keras/deeplabv3plus/keras/deeplab_v3_plus_resnet50_pascalvoc/3", | ||
| "kaggle://keras/densenet/keras/densenet_121_imagenet/2", | ||
| "kaggle://keras/densenet/keras/densenet_169_imagenet/2", | ||
| "kaggle://keras/densenet/keras/densenet_201_imagenet/2", | ||
| "kaggle://keras/mit/keras/mit_b0_ade20k_512/1", | ||
| "kaggle://keras/mit/keras/mit_b1_ade20k_512/1", | ||
| "kaggle://keras/mit/keras/mit_b2_ade20k_512/1", | ||
| "kaggle://keras/mit/keras/mit_b3_ade20k_512/1", | ||
| "kaggle://keras/mit/keras/mit_b4_ade20k_512/1", | ||
| "kaggle://keras/mit/keras/mit_b0_cityscapes_1024/1", | ||
| "kaggle://keras/mit/keras/mit_b1_cityscapes_1024/1", | ||
| "kaggle://keras/mit/keras/mit_b2_cityscapes_1024/1", | ||
| "kaggle://keras/mit/keras/mit_b3_cityscapes_1024/1", | ||
| "kaggle://keras/mit/keras/mit_b4_cityscapes_1024/1", | ||
| "kaggle://keras/mit/keras/mit_b5_cityscapes_1024/1", | ||
| "kaggle://keras/gemma/keras/gemma_2b_en/2", | ||
| "kaggle://keras/gemma/keras/gemma_instruct_2b_en/2", | ||
| "kaggle://keras/gemma/keras/gemma_1.1_instruct_2b_en/3", | ||
| "kaggle://keras/codegemma/keras/code_gemma_1.1_2b_en/1", | ||
| "kaggle://keras/codegemma/keras/code_gemma_2b_en/1", | ||
| "kaggle://keras/gemma/keras/gemma_7b_en/2", | ||
| "kaggle://keras/gemma/keras/gemma_instruct_7b_en/2", | ||
| "kaggle://keras/gemma/keras/gemma_1.1_instruct_7b_en/3", | ||
| "kaggle://keras/codegemma/keras/code_gemma_7b_en/1", | ||
| "kaggle://keras/codegemma/keras/code_gemma_instruct_7b_en/1", | ||
| "kaggle://keras/codegemma/keras/code_gemma_1.1_instruct_7b_en/1", | ||
| "kaggle://keras/gemma2/keras/gemma2_2b_en/1", | ||
| "kaggle://keras/gemma2/keras/gemma2_instruct_2b_en/1", | ||
| "kaggle://keras/gemma2/keras/gemma2_9b_en/2", | ||
| "kaggle://keras/gemma2/keras/gemma2_instruct_9b_en/2", | ||
| "kaggle://keras/gemma2/keras/gemma2_27b_en/1", | ||
| "kaggle://keras/gemma2/keras/gemma2_instruct_27b_en/1", | ||
| "kaggle://google/shieldgemma/keras/shieldgemma_2b_en/1", | ||
| "kaggle://google/shieldgemma/keras/shieldgemma_9b_en/1", | ||
| "kaggle://google/shieldgemma/keras/shieldgemma_27b_en/1", | ||
| "kaggle://keras/paligemma/keras/pali_gemma_3b_mix_224/3", | ||
| "kaggle://keras/paligemma/keras/pali_gemma_3b_mix_448/3", | ||
| "kaggle://keras/paligemma/keras/pali_gemma_3b_224/3", | ||
| "kaggle://keras/paligemma/keras/pali_gemma_3b_448/3", | ||
| "kaggle://keras/paligemma/keras/pali_gemma_3b_896/3", | ||
| "kaggle://keras/resnetv1/keras/resnet_18_imagenet/2", | ||
| "kaggle://keras/resnetv1/keras/resnet_50_imagenet/2", | ||
| "kaggle://keras/resnetv1/keras/resnet_101_imagenet/2", | ||
| "kaggle://keras/resnetv1/keras/resnet_152_imagenet/2", | ||
| "kaggle://keras/resnetv2/keras/resnet_v2_50_imagenet/2", | ||
| "kaggle://keras/resnetv2/keras/resnet_v2_101_imagenet/2", | ||
| "kaggle://keras/sam/keras/sam_base_sa1b/4", | ||
| "kaggle://keras/sam/keras/sam_large_sa1b/4", | ||
| "kaggle://keras/sam/keras/sam_huge_sa1b/4", | ||
| "kaggle://kerashub/segformer/keras/segformer_b0_ade20k_512", | ||
| "kaggle://kerashub/segformer/keras/segformer_b1_ade20k_512", | ||
| "kaggle://kerashub/segformer/keras/segformer_b2_ade20k_512", | ||
| "kaggle://kerashub/segformer/keras/segformer_b3_ade20k_512", | ||
| "kaggle://kerashub/segformer/keras/segformer_b4_ade20k_512", | ||
| "kaggle://kerashub/segformer/keras/segformer_b5_ade20k_640", | ||
| "kaggle://kerashub/segformer/keras/segformer_b0_cityscapes_1024", | ||
| "kaggle://kerashub/segformer/keras/segformer_b1_ade20k_512", | ||
| "kaggle://kerashub/segformer/keras/segformer_b2_cityscapes_1024", | ||
| "kaggle://kerashub/segformer/keras/segformer_b3_cityscapes_1024", | ||
| "kaggle://kerashub/segformer/keras/segformer_b4_cityscapes_1024", | ||
| "kaggle://kerashub/segformer/keras/segformer_b5_cityscapes_1024", | ||
| "kaggle://keras/vgg/keras/vgg_11_imagenet/1", | ||
| "kaggle://keras/vgg/keras/vgg_13_imagenet/1", | ||
| "kaggle://keras/vgg/keras/vgg_16_imagenet/1", | ||
| "kaggle://keras/vgg/keras/vgg_19_imagenet/1", | ||
| "kaggle://keras/whisper/keras/whisper_tiny_en/3", | ||
| "kaggle://keras/whisper/keras/whisper_base_en/3", | ||
| "kaggle://keras/whisper/keras/whisper_small_en/3", | ||
| "kaggle://keras/whisper/keras/whisper_medium_en/3", | ||
| "kaggle://keras/whisper/keras/whisper_tiny_multi/3", | ||
| "kaggle://keras/whisper/keras/whisper_base_multi/3", | ||
| "kaggle://keras/whisper/keras/whisper_small_multi/3", | ||
| "kaggle://keras/whisper/keras/whisper_medium_multi/3", | ||
| "kaggle://keras/whisper/keras/whisper_large_multi/3", | ||
| "kaggle://keras/whisper/keras/whisper_large_multi_v2/3", | ||
| "kaggle://keras/albert/keras/albert_base_en_uncased/2", | ||
| "kaggle://keras/albert/keras/albert_large_en_uncased/2", | ||
| "kaggle://keras/albert/keras/albert_extra_large_en_uncased/2", | ||
| "kaggle://keras/albert/keras/albert_extra_extra_large_en_uncased/2", | ||
| "kaggle://keras/bart/keras/bart_base_en/2", | ||
| "kaggle://keras/bart/keras/bart_large_en/2", | ||
| "kaggle://keras/bart/keras/bart_large_en_cnn/2", | ||
| "kaggle://keras/bert/keras/bert_tiny_en_uncased/2", | ||
| "kaggle://keras/bert/keras/bert_small_en_uncased/2", | ||
| "kaggle://keras/bert/keras/bert_medium_en_uncased/2", | ||
| "kaggle://keras/bert/keras/bert_base_en_uncased/2", | ||
| "kaggle://keras/bert/keras/bert_base_en/2", | ||
| "kaggle://keras/bert/keras/bert_base_zh/2", | ||
| "kaggle://keras/bert/keras/bert_base_multi/2", | ||
| "kaggle://keras/bert/keras/bert_large_en_uncased/2", | ||
| "kaggle://keras/bert/keras/bert_large_en/2", | ||
| "kaggle://keras/bert/keras/bert_tiny_en_uncased_sst2/4", | ||
| "kaggle://keras/bloom/keras/bloom_560m_multi/3", | ||
| "kaggle://keras/bloom/keras/bloom_1.1b_multi/1", | ||
| "kaggle://keras/bloom/keras/bloom_1.7b_multi/1", | ||
| "kaggle://keras/bloom/keras/bloom_3b_multi/1", | ||
| "kaggle://keras/bloom/keras/bloomz_560m_multi/1", | ||
| "kaggle://keras/bloom/keras/bloomz_1.1b_multi/1", | ||
| "kaggle://keras/bloom/keras/bloomz_1.7b_multi/1", | ||
| "kaggle://keras/bloom/keras/bloomz_3b_multi/1", | ||
| "kaggle://keras/deberta_v3/keras/deberta_v3_extra_small_en/2", | ||
| "kaggle://keras/deberta_v3/keras/deberta_v3_small_en/2", | ||
| "kaggle://keras/deberta_v3/keras/deberta_v3_base_en/2", | ||
| "kaggle://keras/deberta_v3/keras/deberta_v3_large_en/2", | ||
| "kaggle://keras/deberta_v3/keras/deberta_v3_base_multi/2", | ||
| "kaggle://keras/distil_bert/keras/distil_bert_base_en_uncased/2", | ||
| "kaggle://keras/distil_bert/keras/distil_bert_base_en/2", | ||
| "kaggle://keras/distil_bert/keras/distil_bert_base_multi/2", | ||
| "kaggle://keras/electra/keras/electra_small_discriminator_uncased_en/1", | ||
| "kaggle://keras/electra/keras/electra_small_generator_uncased_en/1", | ||
| "kaggle://keras/electra/keras/electra_base_discriminator_uncased_en/1", | ||
| "kaggle://keras/electra/keras/electra_base_generator_uncased_en/1", | ||
| "kaggle://keras/electra/keras/electra_large_discriminator_uncased_en/1", | ||
| "kaggle://keras/electra/keras/electra_large_generator_uncased_en/1", | ||
| "kaggle://keras/f_net/keras/f_net_base_en/2", | ||
| "kaggle://keras/f_net/keras/f_net_large_en/2", | ||
| "kaggle://keras/falcon/keras/falcon_refinedweb_1b_en/1", | ||
| "kaggle://keras/gpt2/keras/gpt2_base_en/2", | ||
| "kaggle://keras/gpt2/keras/gpt2_medium_en/2", | ||
| "kaggle://keras/gpt2/keras/gpt2_large_en/2", | ||
| "kaggle://keras/gpt2/keras/gpt2_extra_large_en/2", | ||
| "kaggle://keras/gpt2/keras/gpt2_base_en_cnn_dailymail/2", | ||
| "kaggle://keras/llama2/keras/llama2_7b_en/1", | ||
| "kaggle://keras/llama2/keras/llama2_7b_en_int8/1", | ||
| "kaggle://keras/llama2/keras/llama2_instruct_7b_en/1", | ||
| "kaggle://keras/llama2/keras/llama2_instruct_7b_en_int8/1", | ||
| "kaggle://keras/vicuna/keras/vicuna_1.5_7b_en/1", | ||
| "kaggle://keras/mistral/keras/mistral_7b_en/6", | ||
| "kaggle://keras/mistral/keras/mistral_instruct_7b_en/6", | ||
| "kaggle://keras/mistral/keras/mistral_0.2_instruct_7b_en/1", | ||
| "kaggle://keras/opt/keras/opt_125m_en/2", | ||
| "kaggle://keras/opt/keras/opt_1.3b_en/2", | ||
| "kaggle://keras/opt/keras/opt_2.7b_en/2", | ||
| "kaggle://keras/opt/keras/opt_6.7b_en/2", | ||
| "kaggle://keras/phi3/keras/phi3_mini_4k_instruct_en", | ||
| "kaggle://keras/phi3/keras/phi3_mini_128k_instruct_en", | ||
| "kaggle://keras/roberta/keras/roberta_base_en/2", | ||
| "kaggle://keras/roberta/keras/roberta_large_en/2", | ||
| "kaggle://keras/stablediffusion3/keras/stable_diffusion_3_medium/3", | ||
| "kaggle://keras/t5/keras/t5_small_multi/2", | ||
| "kaggle://keras/t5/keras/t5_base_multi/2", | ||
| "kaggle://keras/t5/keras/t5_large_multi/2", | ||
| "kaggle://keras/t5/keras/flan_small_multi/2", | ||
| "kaggle://keras/t5/keras/flan_base_multi/2", | ||
| "kaggle://keras/t5/keras/flan_large_multi/2", | ||
| "kaggle://keras/xlm_roberta/keras/xlm_roberta_base_multi/2", | ||
| "kaggle://keras/xlm_roberta/keras/xlm_roberta_large_multi/2", | ||
| ] |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,97 @@ | ||
| import json | ||
| import shutil | ||
|
|
||
| import keras_hub | ||
| import keras_hub.src.utils.preset_utils as utils | ||
|
|
||
| try: | ||
| import kagglehub | ||
| except ImportError: | ||
| kagglehub = None | ||
|
|
||
| HF_BASE_URI = "hf://keras" | ||
| JSON_FILE_PATH = "tools/hf_uploaded_presets.json" | ||
|
|
||
|
|
||
| def load_latest_hf_uploads(json_file_path): | ||
| # Load the latest HF uploads from JSON | ||
| with open(json_file_path, "r") as json_file: | ||
| latest_hf_uploads = set(json.load(json_file)) | ||
| print("Loaded latest HF uploads from JSON file.") | ||
| return latest_hf_uploads | ||
|
|
||
|
|
||
| def download_and_upload_missing_models(missing_in_hf_uploads): | ||
| uploaded_handles = [] | ||
| errored_uploads = [] | ||
| for kaggle_handle in missing_in_hf_uploads: | ||
| try: | ||
| model_variant = kaggle_handle.split("/")[3] | ||
| hf_uri = f"{HF_BASE_URI}/{model_variant}" | ||
| kaggle_handle_path = kaggle_handle.removeprefix("kaggle://") | ||
|
|
||
| # Skip Gemma models | ||
| if "gemma" in kaggle_handle_path: | ||
| print(f"Skipping Gemma model preset: {kaggle_handle_path}") | ||
| continue | ||
|
|
||
| print(f"Downloading model: {kaggle_handle_path}") | ||
| model_file_path = kagglehub.model_download(kaggle_handle_path) | ||
|
|
||
| print(f"Uploading to HF: {hf_uri}") | ||
| keras_hub.upload_preset(hf_uri, model_file_path) | ||
|
|
||
| print(f"Cleaning up: {model_file_path}") | ||
| shutil.rmtree(model_file_path) | ||
|
|
||
| # Add to the list of successfully uploaded handles | ||
| uploaded_handles.append(kaggle_handle) | ||
| except Exception as e: | ||
| print( | ||
| f"Error in downloading and uploading preset {kaggle_handle}: {e}" | ||
| ) | ||
| errored_uploads.append(kaggle_handle) | ||
|
|
||
| print("All missing models processed.") | ||
| return uploaded_handles, errored_uploads | ||
|
|
||
|
|
||
| def update_hf_uploads_json(json_file_path, latest_kaggle_handles): | ||
| with open(json_file_path, "w") as json_file: | ||
| json.dump(latest_kaggle_handles, json_file, indent=4) | ||
|
|
||
| print("Updated hf_uploaded_presets.json with newly uploaded handles.") | ||
|
|
||
|
|
||
| def main(): | ||
| print("Starting the model presets mirroring on HF") | ||
|
|
||
| # Step 1: Load presets | ||
| presets = utils.BUILTIN_PRESETS | ||
| print("Loaded presets from utils.") | ||
|
|
||
| # Step 2: Load latest HF uploads | ||
| latest_hf_uploads = load_latest_hf_uploads(JSON_FILE_PATH) | ||
|
|
||
| # Step 3: Find missing uploads | ||
| latest_kaggle_handles = { | ||
| data["kaggle_handle"] for model, data in presets.items() | ||
| } | ||
| missing_in_hf_uploads = latest_kaggle_handles - latest_hf_uploads | ||
| print(f"Found {len(missing_in_hf_uploads)} models missing on HF.") | ||
|
|
||
| # Step 4: Download and upload missing models | ||
| _, errored_uploads = download_and_upload_missing_models( | ||
| missing_in_hf_uploads | ||
| ) | ||
|
|
||
| # Step 5: Update JSON file with newly uploaded handles | ||
| update_hf_uploads_json( | ||
| JSON_FILE_PATH, {latest_kaggle_handles} - {errored_uploads} | ||
| ) | ||
| print("uploads for the following models failed: ", errored_uploads) | ||
| print("Rest of the models up to date on HuggingFace") | ||
|
|
||
|
|
||
| if __name__ == "__main__": | ||
| main() |
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Are we planning to run this script automatically?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It is a good idea to set it up to run automatically once a day. For now it is run manually.