Upload config.json with huggingface_hub
Browse files- config.json +32 -0
config.json
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"MiniMindLM"
|
| 4 |
+
],
|
| 5 |
+
"auto_map": {
|
| 6 |
+
"AutoConfig": "LMConfig.LMConfig",
|
| 7 |
+
"AutoModelForCausalLM": "model.MiniMindLM"
|
| 8 |
+
},
|
| 9 |
+
"aux_loss_alpha": 0.1,
|
| 10 |
+
"dim": 512,
|
| 11 |
+
"dropout": 0.0,
|
| 12 |
+
"flash_attn": true,
|
| 13 |
+
"hidden_dim": 1408,
|
| 14 |
+
"max_seq_len": 8192,
|
| 15 |
+
"model_type": "minimind",
|
| 16 |
+
"multiple_of": 64,
|
| 17 |
+
"n_heads": 8,
|
| 18 |
+
"n_kv_heads": 2,
|
| 19 |
+
"n_layers": 8,
|
| 20 |
+
"n_routed_experts": 4,
|
| 21 |
+
"n_shared_experts": true,
|
| 22 |
+
"norm_eps": 1e-05,
|
| 23 |
+
"norm_topk_prob": true,
|
| 24 |
+
"num_experts_per_tok": 2,
|
| 25 |
+
"rope_theta": 1000000.0,
|
| 26 |
+
"scoring_func": "softmax",
|
| 27 |
+
"seq_aux": true,
|
| 28 |
+
"torch_dtype": "float32",
|
| 29 |
+
"transformers_version": "4.48.0",
|
| 30 |
+
"use_moe": false,
|
| 31 |
+
"vocab_size": 6400
|
| 32 |
+
}
|