commit files to HF hub
Browse files- README.md +25 -0
- config.json +34 -0
- inference.py +10 -0
- openvino_model.bin +3 -0
- openvino_model.xml +0 -0
- special_tokens_map.json +30 -0
- tokenizer.model +3 -0
- tokenizer_config.json +44 -0
README.md
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
tags:
|
5 |
+
- openvino
|
6 |
+
---
|
7 |
+
|
8 |
+
# baichuan-inc/Baichuan2-13B-Chat
|
9 |
+
|
10 |
+
This is the [baichuan-inc/Baichuan2-13B-Chat](https://huggingface.co/baichuan-inc/Baichuan2-13B-Chat) model converted to [OpenVINO](https://openvino.ai) with INT8 weights compression for accelerated inference.
|
11 |
+
|
12 |
+
An example of how to do inference on this model:
|
13 |
+
```python
|
14 |
+
from optimum.intel import OVModelForCausalLM
|
15 |
+
from transformers import AutoTokenizer, pipeline
|
16 |
+
|
17 |
+
# model_id should be set to either a local directory or a model available on the HuggingFace hub.
|
18 |
+
model_id = "helenai/baichuan-inc-Baichuan2-13B-Chat-ov"
|
19 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
|
20 |
+
model = OVModelForCausalLM.from_pretrained(model_id)
|
21 |
+
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
|
22 |
+
result = pipe("hello world")
|
23 |
+
print(result)
|
24 |
+
```
|
25 |
+
|
config.json
ADDED
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"_name_or_path": "baichuan-inc/Baichuan2-13B-Chat",
|
4 |
+
"architectures": [
|
5 |
+
"BaichuanForCausalLM"
|
6 |
+
],
|
7 |
+
"auto_map": {
|
8 |
+
"AutoConfig": "baichuan-inc/Baichuan2-13B-Chat--configuration_baichuan.BaichuanConfig",
|
9 |
+
"AutoModelForCausalLM": "baichuan-inc/Baichuan2-13B-Chat--modeling_baichuan.BaichuanForCausalLM"
|
10 |
+
},
|
11 |
+
"bos_token_id": 1,
|
12 |
+
"eos_token_id": 2,
|
13 |
+
"gradient_checkpointing": [
|
14 |
+
false
|
15 |
+
],
|
16 |
+
"hidden_act": "silu",
|
17 |
+
"hidden_size": 5120,
|
18 |
+
"initializer_range": 0.02,
|
19 |
+
"intermediate_size": 13696,
|
20 |
+
"is_decoder": true,
|
21 |
+
"model_max_length": 4096,
|
22 |
+
"model_type": "baichuan",
|
23 |
+
"num_attention_heads": 40,
|
24 |
+
"num_hidden_layers": 40,
|
25 |
+
"pad_token_id": 0,
|
26 |
+
"rms_norm_eps": 1e-06,
|
27 |
+
"tie_word_embeddings": false,
|
28 |
+
"tokenizer_class": "BaichuanTokenizer",
|
29 |
+
"torch_dtype": "bfloat16",
|
30 |
+
"transformers_version": "4.39.0",
|
31 |
+
"use_cache": true,
|
32 |
+
"vocab_size": 125696,
|
33 |
+
"z_loss_weight": 0
|
34 |
+
}
|
inference.py
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from optimum.intel import OVModelForCausalLM
|
2 |
+
from transformers import AutoTokenizer, pipeline
|
3 |
+
|
4 |
+
# model_id should be set to either a local directory or a model available on the HuggingFace hub.
|
5 |
+
model_id = "helenai/baichuan-inc-Baichuan2-13B-Chat-ov"
|
6 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
|
7 |
+
model = OVModelForCausalLM.from_pretrained(model_id)
|
8 |
+
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
|
9 |
+
result = pipe("hello world")
|
10 |
+
print(result)
|
openvino_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:25cb2ae0f6439944e7432d7441e49253ed7f106a8fc3832a40a9d02107bfdc10
|
3 |
+
size 16589380381
|
openvino_model.xml
ADDED
The diff for this file is too large to render.
See raw diff
|
|
special_tokens_map.json
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": true,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": true
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": true,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": true
|
15 |
+
},
|
16 |
+
"pad_token": {
|
17 |
+
"content": "<unk>",
|
18 |
+
"lstrip": false,
|
19 |
+
"normalized": true,
|
20 |
+
"rstrip": false,
|
21 |
+
"single_word": true
|
22 |
+
},
|
23 |
+
"unk_token": {
|
24 |
+
"content": "<unk>",
|
25 |
+
"lstrip": false,
|
26 |
+
"normalized": true,
|
27 |
+
"rstrip": false,
|
28 |
+
"single_word": true
|
29 |
+
}
|
30 |
+
}
|
tokenizer.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:79452955be6b419a65984273a9f08af86042e1c2a75ee3ba989cbf620a133cc2
|
3 |
+
size 2001107
|
tokenizer_config.json
ADDED
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_bos_token": false,
|
3 |
+
"add_eos_token": false,
|
4 |
+
"added_tokens_decoder": {
|
5 |
+
"0": {
|
6 |
+
"content": "<unk>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": true,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": true,
|
11 |
+
"special": true
|
12 |
+
},
|
13 |
+
"1": {
|
14 |
+
"content": "<s>",
|
15 |
+
"lstrip": false,
|
16 |
+
"normalized": true,
|
17 |
+
"rstrip": false,
|
18 |
+
"single_word": true,
|
19 |
+
"special": true
|
20 |
+
},
|
21 |
+
"2": {
|
22 |
+
"content": "</s>",
|
23 |
+
"lstrip": false,
|
24 |
+
"normalized": true,
|
25 |
+
"rstrip": false,
|
26 |
+
"single_word": true,
|
27 |
+
"special": true
|
28 |
+
}
|
29 |
+
},
|
30 |
+
"auto_map": {
|
31 |
+
"AutoTokenizer": [
|
32 |
+
"baichuan-inc/Baichuan2-13B-Chat--tokenization_baichuan.BaichuanTokenizer",
|
33 |
+
null
|
34 |
+
]
|
35 |
+
},
|
36 |
+
"bos_token": "<s>",
|
37 |
+
"clean_up_tokenization_spaces": false,
|
38 |
+
"eos_token": "</s>",
|
39 |
+
"model_max_length": 4096,
|
40 |
+
"pad_token": "<unk>",
|
41 |
+
"sp_model_kwargs": {},
|
42 |
+
"tokenizer_class": "BaichuanTokenizer",
|
43 |
+
"unk_token": "<unk>"
|
44 |
+
}
|