Commit
•
1ac088c
1
Parent(s):
b144159
Upload folder using huggingface_hub
Browse files- README.md +63 -0
- added_tokens.json +5 -0
- config.json +27 -0
- generation_config.json +9 -0
- gptq_model-4bit-128g.safetensors +3 -0
- quantize_config.json +11 -0
- special_tokens_map.json +27 -0
- tokenizer.json +0 -0
- tokenizer_config.json +43 -0
- vocab.json +0 -0
README.md
ADDED
@@ -0,0 +1,63 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
library_name: transformers
|
3 |
+
license: other
|
4 |
+
datasets:
|
5 |
+
- teknium/OpenHermes-2.5
|
6 |
+
- LDJnr/Capybara
|
7 |
+
- Intel/orca_dpo_pairs
|
8 |
+
- argilla/distilabel-capybara-dpo-7k-binarized
|
9 |
+
language:
|
10 |
+
- en
|
11 |
+
pipeline_tag: text-generation
|
12 |
+
---
|
13 |
+
|
14 |
+
# Quyen
|
15 |
+
<img src="quyen.webp" width="512" height="512" alt="Quyen">
|
16 |
+
|
17 |
+
# Model Description
|
18 |
+
Quyen is our first flagship LLM series based on the Qwen1.5 family. We introduced 6 different versions:
|
19 |
+
|
20 |
+
- **Quyen-SE (0.5B)**
|
21 |
+
- **Quyen-Mini (1.8B)**
|
22 |
+
- **Quyen (4B)**
|
23 |
+
- **Quyen-Plus (7B)**
|
24 |
+
- **Quyen-Pro (14B)**
|
25 |
+
- **Quyen-Pro-Max (72B)**
|
26 |
+
|
27 |
+
All models were trained with SFT and DPO using the following dataset:
|
28 |
+
|
29 |
+
- *OpenHermes-2.5* by **Teknium**
|
30 |
+
- *Capyabara* by **LDJ**
|
31 |
+
- *argilla/distilabel-capybara-dpo-7k-binarized* by **argilla**
|
32 |
+
- *orca_dpo_pairs* by **Intel**
|
33 |
+
- and Private Data by **Ontocord** & **BEE-spoke-data**
|
34 |
+
|
35 |
+
# Prompt Template
|
36 |
+
- All Quyen models use ChatML as the default template:
|
37 |
+
|
38 |
+
```
|
39 |
+
<|im_start|>system
|
40 |
+
You are a sentient, superintelligent artificial general intelligence, here to teach and assist me.<|im_end|>
|
41 |
+
<|im_start|>user
|
42 |
+
Hello world.<|im_end|>
|
43 |
+
<|im_start|>assistant
|
44 |
+
```
|
45 |
+
|
46 |
+
- You can also use `apply_chat_template`:
|
47 |
+
|
48 |
+
```python
|
49 |
+
messages = [
|
50 |
+
{"role": "system", "content": "You are a sentient, superintelligent artificial general intelligence, here to teach and assist me."},
|
51 |
+
{"role": "user", "content": "Hello world."}
|
52 |
+
]
|
53 |
+
gen_input = tokenizer.apply_chat_template(message, return_tensors="pt")
|
54 |
+
model.generate(**gen_input)
|
55 |
+
```
|
56 |
+
|
57 |
+
# Benchmarks:
|
58 |
+
|
59 |
+
- Coming Soon! We will update the benchmarks later
|
60 |
+
|
61 |
+
# Acknowledgement
|
62 |
+
- We're incredibly grateful to **Tensoic** and **Ontocord** for their generous support with compute and data preparation.
|
63 |
+
- Special thanks to the Qwen team for letting us access the models early for these amazing finetunes.
|
added_tokens.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"<|endoftext|>": 151643,
|
3 |
+
"<|im_end|>": 151645,
|
4 |
+
"<|im_start|>": 151644
|
5 |
+
}
|
config.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "/models/models/Quyen-Mini-v0.1",
|
3 |
+
"architectures": [
|
4 |
+
"Qwen2ForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"eos_token_id": 151645,
|
8 |
+
"hidden_act": "silu",
|
9 |
+
"hidden_size": 2048,
|
10 |
+
"initializer_range": 0.02,
|
11 |
+
"intermediate_size": 5504,
|
12 |
+
"max_position_embeddings": 8192,
|
13 |
+
"max_window_layers": 21,
|
14 |
+
"model_type": "qwen2",
|
15 |
+
"num_attention_heads": 16,
|
16 |
+
"num_hidden_layers": 24,
|
17 |
+
"num_key_value_heads": 16,
|
18 |
+
"rms_norm_eps": 1e-06,
|
19 |
+
"rope_theta": 1000000.0,
|
20 |
+
"sliding_window": 4096,
|
21 |
+
"tie_word_embeddings": false,
|
22 |
+
"torch_dtype": "float16",
|
23 |
+
"transformers_version": "4.37.1",
|
24 |
+
"use_cache": false,
|
25 |
+
"use_sliding_window": false,
|
26 |
+
"vocab_size": 151936
|
27 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token_id": 151643,
|
3 |
+
"do_sample": true,
|
4 |
+
"eos_token_id": 151643,
|
5 |
+
"max_length": 4096,
|
6 |
+
"temperature": 0.7,
|
7 |
+
"top_p": 0.8,
|
8 |
+
"transformers_version": "4.38.0.dev0"
|
9 |
+
}
|
gptq_model-4bit-128g.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f61193a2382c25ca3569db32c833cd47c3e993ed4e9963d62be3aa458b14cd11
|
3 |
+
size 1878528072
|
quantize_config.json
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bits": 4,
|
3 |
+
"group_size": 128,
|
4 |
+
"damp_percent": 0.01,
|
5 |
+
"desc_act": true,
|
6 |
+
"static_groups": false,
|
7 |
+
"sym": true,
|
8 |
+
"true_sequential": true,
|
9 |
+
"model_name_or_path": null,
|
10 |
+
"model_file_base_name": null
|
11 |
+
}
|
special_tokens_map.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"additional_special_tokens": [
|
3 |
+
"<|im_start|>",
|
4 |
+
"<|im_end|>"
|
5 |
+
],
|
6 |
+
"bos_token": {
|
7 |
+
"content": "<|im_end|>",
|
8 |
+
"lstrip": false,
|
9 |
+
"normalized": false,
|
10 |
+
"rstrip": false,
|
11 |
+
"single_word": false
|
12 |
+
},
|
13 |
+
"eos_token": {
|
14 |
+
"content": "<|im_end|>",
|
15 |
+
"lstrip": false,
|
16 |
+
"normalized": false,
|
17 |
+
"rstrip": false,
|
18 |
+
"single_word": false
|
19 |
+
},
|
20 |
+
"pad_token": {
|
21 |
+
"content": "<|endoftext|>",
|
22 |
+
"lstrip": false,
|
23 |
+
"normalized": false,
|
24 |
+
"rstrip": false,
|
25 |
+
"single_word": false
|
26 |
+
}
|
27 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"added_tokens_decoder": {
|
4 |
+
"151643": {
|
5 |
+
"content": "<|endoftext|>",
|
6 |
+
"lstrip": false,
|
7 |
+
"normalized": false,
|
8 |
+
"rstrip": false,
|
9 |
+
"single_word": false,
|
10 |
+
"special": true
|
11 |
+
},
|
12 |
+
"151644": {
|
13 |
+
"content": "<|im_start|>",
|
14 |
+
"lstrip": false,
|
15 |
+
"normalized": false,
|
16 |
+
"rstrip": false,
|
17 |
+
"single_word": false,
|
18 |
+
"special": false
|
19 |
+
},
|
20 |
+
"151645": {
|
21 |
+
"content": "<|im_end|>",
|
22 |
+
"lstrip": false,
|
23 |
+
"normalized": false,
|
24 |
+
"rstrip": false,
|
25 |
+
"single_word": false,
|
26 |
+
"special": true
|
27 |
+
}
|
28 |
+
},
|
29 |
+
"additional_special_tokens": [
|
30 |
+
"<|im_start|>",
|
31 |
+
"<|im_end|>"
|
32 |
+
],
|
33 |
+
"bos_token": "<|im_end|>",
|
34 |
+
"chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
|
35 |
+
"clean_up_tokenization_spaces": false,
|
36 |
+
"eos_token": "<|im_end|>",
|
37 |
+
"errors": "replace",
|
38 |
+
"model_max_length": 32768,
|
39 |
+
"pad_token": "<|endoftext|>",
|
40 |
+
"split_special_tokens": false,
|
41 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
42 |
+
"unk_token": null
|
43 |
+
}
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|