| { |
| "architectures": [ |
| "AriaForConditionalGeneration" |
| ], |
| "auto_map": { |
| "AutoConfig": "modeling_aria.AriaConfig", |
| "AutoModelForCausalLM": "modeling_aria.AriaForConditionalGeneration" |
| }, |
| "ignore_index": -100, |
| "image_token_index": 9, |
| "model_type": "aria", |
| "projector_patch_to_query_dict": { |
| "1225": 128, |
| "4900": 256 |
| }, |
| "text_config": { |
| "hidden_size": 2560, |
| "intermediate_size": 13568, |
| "max_position_embeddings": 65536, |
| "model_type": "aria_moe_lm", |
| "moe_intermediate_size": 1664, |
| "moe_num_experts": 64, |
| "moe_topk": 6, |
| "num_attention_heads": 20, |
| "num_experts_per_tok": 6, |
| "num_hidden_layers": 28, |
| "num_key_value_heads": 20, |
| "rope_theta": 5000000, |
| "vocab_size": 100352 |
| }, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.45.0", |
| "_attn_implementation": "flash_attention_2", |
| "vision_config": { |
| "_flash_attn_2_enabled": true, |
| "_attn_implementation": "flash_attention_2", |
| "architectures": [ |
| "AriaVisionModel" |
| ], |
| "hidden_size": 1152, |
| "image_size": 980, |
| "intermediate_size": 4304, |
| "model_type": "aria_vision_model", |
| "num_attention_heads": 16, |
| "num_hidden_layers": 27, |
| "patch_size": 14, |
| "torch_dtype": "bfloat16" |
| } |
| } |
|
|