-
Link directly to Dataset encoder data
-
Keep the model card as a reference and link to GitHub workflows
-
Connect the model card to the encoder code and first-run workflow
-
Organize the model guide around use, inputs, and reproduction
-
Organize the model guide around use, inputs, and reproduction
-
Provide complete training recipes and the reporting protocol
-
Use the published HF repositories directly
-
Use the published GitHub and Dataset paths
-
Clarify the released 294-step input contract
-
Add masked-reconstruction evaluation results
-
Complete the public pretraining recipe
-
Remove legacy training checkpoint files
-
Publish standalone Safetensors encoder artifact
-
Simplify the standalone encoder card
-
Document supported EGMS data workflows
-
Upload folder using huggingface_hub
-
Document dataset-authoritative normalization
-
Add files using upload-large-folder tool
-
Update dataset release paths
-
Flatten encoder repo: encoder.pt + sidecars at root
-
Add files using upload-large-folder tool
-
Delete files encoder.pt encoder_tokens_10k.pt encoder_tokens_10k_metadata.json manifest/split.parquet manifest/data_config.json with huggingface_hub
-
Upload README.md with huggingface_hub
-
Upload README.md with huggingface_hub
-
Add files using upload-large-folder tool
-
initial commit
