| { |
| "_name_or_path": "deepseek-ai/deepseek-coder-1.3b-base", |
| "architectures": [ |
| "DetikzifyForCausalLM" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 32013, |
| "concat_patches": 2, |
| "eos_token_id": 32014, |
| "feature_layer": 26, |
| "hidden_act": "silu", |
| "hidden_size": 2048, |
| "initializer_range": 0.02, |
| "intermediate_size": 5504, |
| "max_position_embeddings": 16384, |
| "mlp_bias": false, |
| "mm_hidden_size": 2304, |
| "model_type": "detikzify", |
| "num_attention_heads": 16, |
| "num_hidden_layers": 24, |
| "num_key_value_heads": 16, |
| "num_patches": 364, |
| "patch_token_id": 32013, |
| "pretraining_tp": 1, |
| "rms_norm_eps": 1e-06, |
| "rope_scaling": { |
| "factor": 4.0, |
| "rope_type": "linear", |
| "type": "linear" |
| }, |
| "rope_theta": 100000, |
| "tie_word_embeddings": false, |
| "torch_dtype": "float32", |
| "transformers_version": "4.44.2", |
| "use_cache": false, |
| "use_mm_proj": true, |
| "vision_config": { |
| "architecture": "vit_so400m_patch14_siglip_384", |
| "classifier": "head", |
| "crop_mode": "center", |
| "crop_pct": 0.9, |
| "custom_load": false, |
| "first_conv": "patch_embed.proj", |
| "fixed_input_size": true, |
| "hf_hub_filename": "open_clip_pytorch_model.bin", |
| "hf_hub_id": "timm/ViT-SO400M-14-SigLIP-384", |
| "input_size": [ |
| 3, |
| 384, |
| 384 |
| ], |
| "interpolation": "bicubic", |
| "mean": [ |
| 0.5, |
| 0.5, |
| 0.5 |
| ], |
| "num_classes": 0, |
| "pool_size": null, |
| "std": [ |
| 0.5, |
| 0.5, |
| 0.5 |
| ], |
| "tag": "webli", |
| "url": "" |
| }, |
| "vision_tower": "vit_so400m_patch14_siglip_384.webli", |
| "vocab_size": 32256 |
| } |
|
|