| { | |
| "model_type": "slip", | |
| "architectures": [ | |
| "SlipForCausalLM" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_slip.SlipConfig", | |
| "AutoModelForCausalLM": "modeling_slip.SlipForCausalLM" | |
| }, | |
| "llm_model_name": "meta-llama/Llama-3.2-1B", | |
| "split_layer": 12, | |
| "num_img_queries": 64, | |
| "num_heads": 32, | |
| "img_attn_pool_num_heads": 8, | |
| "hidden_size": 2048, | |
| "post_train": false, | |
| "max_llm_len": 768, | |
| "vocab_size": 128256, | |
| "pad_token_id": null, | |
| "bos_token_id": 128000, | |
| "eos_token_id": 128001, | |
| "sensor_encoder": { | |
| "embed_dim": 768, | |
| "num_heads": 12, | |
| "mlp_ratio": 4, | |
| "depth": 12, | |
| "dropout_rate": 0.1, | |
| "learnable_pos_emb": false, | |
| "max_position_embeddings": 4880, | |
| "patch_size": null, | |
| "channel_attn_type": "all_attn" | |
| }, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.57.6" | |
| } | |