Upload folder using huggingface_hub
Browse files- config.json +28 -0
- pytorch_model.bin +3 -0
config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"MiniMindLM"
|
4 |
+
],
|
5 |
+
"aux_loss_alpha": 0.1,
|
6 |
+
"dim": 512,
|
7 |
+
"dropout": 0.0,
|
8 |
+
"flash_attn": true,
|
9 |
+
"hidden_dim": 1408,
|
10 |
+
"max_seq_len": 8192,
|
11 |
+
"model_type": "minimind",
|
12 |
+
"multiple_of": 64,
|
13 |
+
"n_heads": 8,
|
14 |
+
"n_kv_heads": 2,
|
15 |
+
"n_layers": 8,
|
16 |
+
"n_routed_experts": 4,
|
17 |
+
"n_shared_experts": true,
|
18 |
+
"norm_eps": 1e-05,
|
19 |
+
"norm_topk_prob": true,
|
20 |
+
"num_experts_per_tok": 2,
|
21 |
+
"rope_theta": 1000000.0,
|
22 |
+
"scoring_func": "softmax",
|
23 |
+
"seq_aux": true,
|
24 |
+
"torch_dtype": "float32",
|
25 |
+
"transformers_version": "4.50.2",
|
26 |
+
"use_moe": false,
|
27 |
+
"vocab_size": 6400
|
28 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:81b672336dfba2b0ab3fb6d1c8f187875d663c5936bd307e4350011844597df8
|
3 |
+
size 103342988
|