Llama-2-7b-chat-hf
#1
by
zhuangwt
- opened
- README.md +0 -57
- mlc-chat-config.json +2 -35
README.md
DELETED
@@ -1,57 +0,0 @@
|
|
1 |
-
---
|
2 |
-
library_name: mlc-llm
|
3 |
-
base_model: meta-llama/Llama-2-7b-chat-hf
|
4 |
-
tags:
|
5 |
-
- mlc-llm
|
6 |
-
- web-llm
|
7 |
-
---
|
8 |
-
|
9 |
-
# Llama-2-7b-chat-hf-q4f16_1-MLC
|
10 |
-
|
11 |
-
This is the [Llama-2-7b-chat-hf](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) model in MLC format `q4f16_1`.
|
12 |
-
The model can be used for projects [MLC-LLM](https://github.com/mlc-ai/mlc-llm) and [WebLLM](https://github.com/mlc-ai/web-llm).
|
13 |
-
|
14 |
-
## Example Usage
|
15 |
-
|
16 |
-
Here are some examples of using this model in MLC LLM.
|
17 |
-
Before running the examples, please install MLC LLM by following the [installation documentation](https://llm.mlc.ai/docs/install/mlc_llm.html#install-mlc-packages).
|
18 |
-
|
19 |
-
### Chat
|
20 |
-
|
21 |
-
In command line, run
|
22 |
-
```bash
|
23 |
-
mlc_llm chat HF://mlc-ai/Llama-2-7b-chat-hf-q4f16_1-MLC
|
24 |
-
```
|
25 |
-
|
26 |
-
### REST Server
|
27 |
-
|
28 |
-
In command line, run
|
29 |
-
```bash
|
30 |
-
mlc_llm serve HF://mlc-ai/Llama-2-7b-chat-hf-q4f16_1-MLC
|
31 |
-
```
|
32 |
-
|
33 |
-
### Python API
|
34 |
-
|
35 |
-
```python
|
36 |
-
from mlc_llm import MLCEngine
|
37 |
-
|
38 |
-
# Create engine
|
39 |
-
model = "HF://mlc-ai/Llama-2-7b-chat-hf-q4f16_1-MLC"
|
40 |
-
engine = MLCEngine(model)
|
41 |
-
|
42 |
-
# Run chat completion in OpenAI API.
|
43 |
-
for response in engine.chat.completions.create(
|
44 |
-
messages=[{"role": "user", "content": "What is the meaning of life?"}],
|
45 |
-
model=model,
|
46 |
-
stream=True,
|
47 |
-
):
|
48 |
-
for choice in response.choices:
|
49 |
-
print(choice.delta.content, end="", flush=True)
|
50 |
-
print("\n")
|
51 |
-
|
52 |
-
engine.terminate()
|
53 |
-
```
|
54 |
-
|
55 |
-
## Documentation
|
56 |
-
|
57 |
-
For more information on MLC LLM project, please visit our [documentation](https://llm.mlc.ai/docs/) and [GitHub repo](http://github.com/mlc-ai/mlc-llm).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
mlc-chat-config.json
CHANGED
@@ -13,8 +13,7 @@
|
|
13 |
"prefill_chunk_size": 4096,
|
14 |
"num_key_value_heads": 32,
|
15 |
"head_dim": 128,
|
16 |
-
"tensor_parallel_shards": 1
|
17 |
-
"max_batch_size": 80
|
18 |
},
|
19 |
"vocab_size": 32000,
|
20 |
"context_window_size": 4096,
|
@@ -28,39 +27,7 @@
|
|
28 |
"temperature": 0.6,
|
29 |
"repetition_penalty": 1.0,
|
30 |
"top_p": 0.9,
|
31 |
-
"conv_template":
|
32 |
-
"name": "llama-2",
|
33 |
-
"system_template": "[INST] <<SYS>>\n{system_message}\n<</SYS>>\n\n",
|
34 |
-
"system_message": "You are a helpful, respectful and honest assistant.",
|
35 |
-
"system_prefix_token_ids": [
|
36 |
-
1
|
37 |
-
],
|
38 |
-
"add_role_after_system_message": false,
|
39 |
-
"roles": {
|
40 |
-
"user": "[INST]",
|
41 |
-
"assistant": "[/INST]",
|
42 |
-
"tool": "[INST]"
|
43 |
-
},
|
44 |
-
"role_templates": {
|
45 |
-
"user": "{user_message}",
|
46 |
-
"assistant": "{assistant_message}",
|
47 |
-
"tool": "{tool_message}"
|
48 |
-
},
|
49 |
-
"messages": [],
|
50 |
-
"seps": [
|
51 |
-
" "
|
52 |
-
],
|
53 |
-
"role_content_sep": " ",
|
54 |
-
"role_empty_sep": " ",
|
55 |
-
"stop_str": [
|
56 |
-
"[INST]"
|
57 |
-
],
|
58 |
-
"stop_token_ids": [
|
59 |
-
2
|
60 |
-
],
|
61 |
-
"function_string": "",
|
62 |
-
"use_function_calling": false
|
63 |
-
},
|
64 |
"pad_token_id": 0,
|
65 |
"bos_token_id": 1,
|
66 |
"eos_token_id": 2,
|
|
|
13 |
"prefill_chunk_size": 4096,
|
14 |
"num_key_value_heads": 32,
|
15 |
"head_dim": 128,
|
16 |
+
"tensor_parallel_shards": 1
|
|
|
17 |
},
|
18 |
"vocab_size": 32000,
|
19 |
"context_window_size": 4096,
|
|
|
27 |
"temperature": 0.6,
|
28 |
"repetition_penalty": 1.0,
|
29 |
"top_p": 0.9,
|
30 |
+
"conv_template": "llama-2",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
31 |
"pad_token_id": 0,
|
32 |
"bos_token_id": 1,
|
33 |
"eos_token_id": 2,
|