| # Core ML stack. torch is provided by the base image. | |
| # TRL >= 0.16 is required for the modern GRPOTrainer / GRPOConfig API. | |
| transformers==4.51.3 | |
| trl==0.16.1 | |
| peft==0.14.0 | |
| accelerate==1.4.0 | |
| bitsandbytes==0.45.5 | |
| datasets==3.3.2 | |
| # Tokenizer + utility deps | |
| sentencepiece==0.2.0 | |
| tiktoken==0.8.0 | |
| einops==0.8.0 | |
| safetensors==0.4.5 | |
| # HTTP + Hub | |
| requests==2.32.3 | |
| huggingface_hub==0.30.2 | |
| # Misc | |
| numpy==1.26.4 | |