Upload BaichuanForCausalLM
Browse files- config.json +28 -0
- generation_config.json +14 -0
- pytorch_model.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"_name_or_path": "baichuan-inc/Baichuan-13B-Chat",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"BaichuanForCausalLM"
|
| 6 |
+
],
|
| 7 |
+
"auto_map": {
|
| 8 |
+
"AutoConfig": "baichuan-inc/Baichuan-13B-Chat--configuration_baichuan.BaichuanConfig",
|
| 9 |
+
"AutoModelForCausalLM": "baichuan-inc/Baichuan-13B-Chat--modeling_baichuan.BaichuanForCausalLM"
|
| 10 |
+
},
|
| 11 |
+
"bos_token_id": 1,
|
| 12 |
+
"eos_token_id": 2,
|
| 13 |
+
"hidden_act": "silu",
|
| 14 |
+
"hidden_size": 5120,
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"intermediate_size": 13696,
|
| 17 |
+
"model_max_length": 4096,
|
| 18 |
+
"model_type": "baichuan",
|
| 19 |
+
"num_attention_heads": 40,
|
| 20 |
+
"num_hidden_layers": 2,
|
| 21 |
+
"pad_token_id": 0,
|
| 22 |
+
"rms_norm_eps": 1e-06,
|
| 23 |
+
"tie_word_embeddings": false,
|
| 24 |
+
"torch_dtype": "float32",
|
| 25 |
+
"transformers_version": "4.31.0",
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_size": 64000
|
| 28 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"assistant_token_id": 196,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"do_sample": true,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"max_new_tokens": 2048,
|
| 7 |
+
"pad_token_id": 0,
|
| 8 |
+
"repetition_penalty": 1.1,
|
| 9 |
+
"temperature": 0.3,
|
| 10 |
+
"top_k": 5,
|
| 11 |
+
"top_p": 0.85,
|
| 12 |
+
"transformers_version": "4.31.0",
|
| 13 |
+
"user_token_id": 195
|
| 14 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb500989c7a63d06eb824ea6785947347535fe73b7a9b2990687023f986ea874
|
| 3 |
+
size 5143373236
|