| { |
| "_name_or_path": "checkpoints/dreamcoder-qwen3_5-0_8b-base-32k/global_step_10000", |
| "architectures": [ |
| "DreamModel" |
| ], |
| "attention_dropout": 0.0, |
| "auto_map": { |
| "AutoConfig": "Dream-org/Dream-Coder-v0-Base-7B--configuration_dream.DreamConfig", |
| "AutoModel": "Dream-org/Dream-Coder-v0-Base-7B--modeling_dream.DreamModel" |
| }, |
| "delete_token_id": 248079, |
| "eos_token_id": 248046, |
| "expand_token_id": 248078, |
| "hidden_act": "silu", |
| "hidden_size": 1024, |
| "initializer_range": 0.02, |
| "intermediate_size": 3584, |
| "mask_token_id": 248077, |
| "max_position_embeddings": 32768, |
| "max_window_layers": 28, |
| "model_type": "Dream", |
| "num_attention_heads": 8, |
| "num_hidden_layers": 24, |
| "num_key_value_heads": 2, |
| "pad_token_id": 248044, |
| "rms_norm_eps": 1e-06, |
| "rope_scaling": null, |
| "rope_theta": 10000000.0, |
| "sliding_window": null, |
| "tie_word_embeddings": true, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.46.2", |
| "use_cache": false, |
| "use_sliding_window": false, |
| "vocab_size": 248320 |
| } |
|
|