Upload config.json with huggingface_hub
Browse files- config.json +49 -0
    	
        config.json
    ADDED
    
    | @@ -0,0 +1,49 @@ | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | |
|  | 
|  | |
| 1 | 
            +
            {
         | 
| 2 | 
            +
              "_name_or_path": "/mnt/bn/ridger1/flame/configs/HGRN2/hgrn2_340M_hybrid_3.json",
         | 
| 3 | 
            +
              "architectures": [
         | 
| 4 | 
            +
                "HGRN2ForCausalLM"
         | 
| 5 | 
            +
              ],
         | 
| 6 | 
            +
              "attn": {
         | 
| 7 | 
            +
                "layers": [
         | 
| 8 | 
            +
                  2,
         | 
| 9 | 
            +
                  5,
         | 
| 10 | 
            +
                  8,
         | 
| 11 | 
            +
                  11,
         | 
| 12 | 
            +
                  14,
         | 
| 13 | 
            +
                  17,
         | 
| 14 | 
            +
                  20,
         | 
| 15 | 
            +
                  23
         | 
| 16 | 
            +
                ],
         | 
| 17 | 
            +
                "num_heads": 16,
         | 
| 18 | 
            +
                "num_kv_heads": 8,
         | 
| 19 | 
            +
                "qkv_bias": false,
         | 
| 20 | 
            +
                "rope_theta": 10000.0,
         | 
| 21 | 
            +
                "window_size": null
         | 
| 22 | 
            +
              },
         | 
| 23 | 
            +
              "attn_mode": "chunk",
         | 
| 24 | 
            +
              "bos_token_id": 1,
         | 
| 25 | 
            +
              "conv_size": 4,
         | 
| 26 | 
            +
              "elementwise_affine": true,
         | 
| 27 | 
            +
              "eos_token_id": 2,
         | 
| 28 | 
            +
              "expand_ratio": 128,
         | 
| 29 | 
            +
              "fuse_cross_entropy": true,
         | 
| 30 | 
            +
              "fuse_norm": true,
         | 
| 31 | 
            +
              "fuse_swiglu": true,
         | 
| 32 | 
            +
              "hidden_act": "swish",
         | 
| 33 | 
            +
              "hidden_ratio": 4,
         | 
| 34 | 
            +
              "hidden_size": 1024,
         | 
| 35 | 
            +
              "initializer_range": 0.02,
         | 
| 36 | 
            +
              "intermediate_size": null,
         | 
| 37 | 
            +
              "max_position_embeddings": 2048,
         | 
| 38 | 
            +
              "model_type": "hgrn2",
         | 
| 39 | 
            +
              "norm_eps": 1e-06,
         | 
| 40 | 
            +
              "num_heads": 8,
         | 
| 41 | 
            +
              "num_hidden_layers": 24,
         | 
| 42 | 
            +
              "tie_word_embeddings": true,
         | 
| 43 | 
            +
              "torch_dtype": "float32",
         | 
| 44 | 
            +
              "transformers_version": "4.49.0",
         | 
| 45 | 
            +
              "use_cache": true,
         | 
| 46 | 
            +
              "use_lower_bound": true,
         | 
| 47 | 
            +
              "use_short_conv": false,
         | 
| 48 | 
            +
              "vocab_size": 32000
         | 
| 49 | 
            +
            }
         | 

