Hello
Browse files- README.md +117 -12
- config.json +40 -0
- generation_config.json +6 -0
- gitattributes +35 -0
- model.safetensors.index.json +0 -0
- output-00001-of-00008.safetensors +3 -0
- output-00002-of-00008.safetensors +3 -0
- output-00003-of-00008.safetensors +3 -0
- output-00004-of-00008.safetensors +3 -0
- output-00005-of-00008.safetensors +3 -0
- output-00006-of-00008.safetensors +3 -0
- output-00007-of-00008.safetensors +3 -0
- output-00008-of-00008.safetensors +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +43 -0
README.md
CHANGED
@@ -1,15 +1,120 @@
|
|
1 |
-
|
|
|
|
|
2 |
|
3 |
-
|
|
|
4 |
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
|
15 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
---
|
4 |
|
5 |
+
# Model Card for Mixtral-8x22B-Instruct-v0.1
|
6 |
+
The Mixtral-8x22B-Instruct-v0.1 Large Language Model (LLM) is an instruct fine-tuned version of the [Mixtral-8x22B-v0.1](https://huggingface.co/mistralai/Mixtral-8x22B-v0.1).
|
7 |
|
8 |
+
## Run the model
|
9 |
+
```python
|
10 |
+
from transformers import AutoModelForCausalLM
|
11 |
+
from mistral_common.protocol.instruct.messages import (
|
12 |
+
AssistantMessage,
|
13 |
+
UserMessage,
|
14 |
+
)
|
15 |
+
from mistral_common.tokens.tokenizers.mistral import MistralTokenizer
|
16 |
+
from mistral_common.tokens.instruct.normalize import ChatCompletionRequest
|
17 |
|
18 |
+
device = "cuda" # the device to load the model onto
|
19 |
+
|
20 |
+
tokenizer_v3 = MistralTokenizer.v3()
|
21 |
+
|
22 |
+
mistral_query = ChatCompletionRequest(
|
23 |
+
tools=[
|
24 |
+
Tool(
|
25 |
+
function=Function(
|
26 |
+
name="get_current_weather",
|
27 |
+
description="Get the current weather",
|
28 |
+
parameters={
|
29 |
+
"type": "object",
|
30 |
+
"properties": {
|
31 |
+
"location": {
|
32 |
+
"type": "string",
|
33 |
+
"description": "The city and state, e.g. San Francisco, CA",
|
34 |
+
},
|
35 |
+
"format": {
|
36 |
+
"type": "string",
|
37 |
+
"enum": ["celsius", "fahrenheit"],
|
38 |
+
"description": "The temperature unit to use. Infer this from the users location.",
|
39 |
+
},
|
40 |
+
},
|
41 |
+
"required": ["location", "format"],
|
42 |
+
},
|
43 |
+
)
|
44 |
+
)
|
45 |
+
],
|
46 |
+
messages=[
|
47 |
+
UserMessage(content="What's the weather like today in Paris"),
|
48 |
+
],
|
49 |
+
model="test",
|
50 |
+
)
|
51 |
+
|
52 |
+
encodeds = tokenizer_v3.encode_chat_completion(mistral_query).tokens
|
53 |
+
model = AutoModelForCausalLM.from_pretrained("mistralai/Mixtral-8x22B-Instruct-v0.1")
|
54 |
+
model_inputs = encodeds.to(device)
|
55 |
+
model.to(device)
|
56 |
+
|
57 |
+
generated_ids = model.generate(model_inputs, max_new_tokens=1000, do_sample=True)
|
58 |
+
sp_tokenizer = tokenizer_v3.instruct_tokenizer.tokenizer
|
59 |
+
decoded = sp_tokenizer.decode(generated_ids[0])
|
60 |
+
print(decoded)
|
61 |
+
```
|
62 |
+
|
63 |
+
# Instruct tokenizer
|
64 |
+
The HuggingFace tokenizer included in this release should match our own. To compare:
|
65 |
+
`pip install mistral-common`
|
66 |
+
|
67 |
+
```py
|
68 |
+
from mistral_common.protocol.instruct.messages import (
|
69 |
+
AssistantMessage,
|
70 |
+
UserMessage,
|
71 |
+
)
|
72 |
+
from mistral_common.tokens.tokenizers.mistral import MistralTokenizer
|
73 |
+
from mistral_common.tokens.instruct.normalize import ChatCompletionRequest
|
74 |
+
|
75 |
+
from transformers import AutoTokenizer
|
76 |
+
|
77 |
+
tokenizer_v3 = MistralTokenizer.v3()
|
78 |
+
|
79 |
+
mistral_query = ChatCompletionRequest(
|
80 |
+
messages=[
|
81 |
+
UserMessage(content="How many experts ?"),
|
82 |
+
AssistantMessage(content="8"),
|
83 |
+
UserMessage(content="How big ?"),
|
84 |
+
AssistantMessage(content="22B"),
|
85 |
+
UserMessage(content="Noice 🎉 !"),
|
86 |
+
],
|
87 |
+
model="test",
|
88 |
+
)
|
89 |
+
hf_messages = mistral_query.model_dump()['messages']
|
90 |
+
|
91 |
+
tokenized_mistral = tokenizer_v3.encode_chat_completion(mistral_query).tokens
|
92 |
+
|
93 |
+
tokenizer_hf = AutoTokenizer.from_pretrained('mistralai/Mixtral-8x22B-Instruct-v0.1')
|
94 |
+
tokenized_hf = tokenizer_hf.apply_chat_template(hf_messages, tokenize=True)
|
95 |
+
|
96 |
+
assert tokenized_hf == tokenized_mistral
|
97 |
+
```
|
98 |
+
|
99 |
+
# Function calling and special tokens
|
100 |
+
This tokenizer includes more special tokens, related to function calling :
|
101 |
+
- [TOOL_CALLS]
|
102 |
+
- [AVAILABLE_TOOLS]
|
103 |
+
- [/AVAILABLE_TOOLS]
|
104 |
+
- [TOOL_RESULT]
|
105 |
+
- [/TOOL_RESULTS]
|
106 |
+
|
107 |
+
If you want to use this model with function calling, please be sure to apply it similarly to what is done in our [SentencePieceTokenizerV3](github.com/mistralai/mistral-common/...).
|
108 |
+
|
109 |
+
# The Mistral AI Team
|
110 |
+
Albert Jiang, Alexandre Sablayrolles, Alexis Tacnet, Antoine Roux,
|
111 |
+
Arthur Mensch, Audrey Herblin-Stoop, Baptiste Bout, Baudouin de Monicault,
|
112 |
+
Blanche Savary, Bam4d, Caroline Feldman, Devendra Singh Chaplot,
|
113 |
+
Diego de las Casas, Eleonore Arcelin, Emma Bou Hanna, Etienne Metzger,
|
114 |
+
Gianna Lengyel, Guillaume Bour, Guillaume Lample, Harizo Rajaona,
|
115 |
+
Jean-Malo Delignon, Jia Li, Justus Murke, Louis Martin, Louis Ternon,
|
116 |
+
Lucile Saulnier, Lélio Renard Lavaud, Margaret Jennings, Marie Pellat,
|
117 |
+
Marie Torelli, Marie-Anne Lachaux, Nicolas Schuhl, Patrick von Platen,
|
118 |
+
Pierre Stock, Sandeep Subramanian, Sophia Yang, Szymon Antoniak, Teven Le Scao,
|
119 |
+
Thibaut Lavril, Timothée Lacroix, Théophile Gervet, Thomas Wang,
|
120 |
+
Valera Nemychnikova, William El Sayed, William Marshall
|
config.json
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"MixtralForCausalLM"
|
4 |
+
],
|
5 |
+
"attention_dropout": 0.0,
|
6 |
+
"bos_token_id": 1,
|
7 |
+
"eos_token_id": 2,
|
8 |
+
"hidden_act": "silu",
|
9 |
+
"hidden_size": 6144,
|
10 |
+
"initializer_range": 0.02,
|
11 |
+
"intermediate_size": 16384,
|
12 |
+
"max_position_embeddings": 65536,
|
13 |
+
"model_type": "mixtral",
|
14 |
+
"num_attention_heads": 48,
|
15 |
+
"num_experts_per_tok": 2,
|
16 |
+
"num_hidden_layers": 56,
|
17 |
+
"num_key_value_heads": 8,
|
18 |
+
"num_local_experts": 8,
|
19 |
+
"output_router_logits": false,
|
20 |
+
"rms_norm_eps": 1e-05,
|
21 |
+
"rope_theta": 1000000.0,
|
22 |
+
"router_aux_loss_coef": 0.001,
|
23 |
+
"sliding_window": null,
|
24 |
+
"tie_word_embeddings": false,
|
25 |
+
"torch_dtype": "bfloat16",
|
26 |
+
"transformers_version": "4.38.0",
|
27 |
+
"use_cache": true,
|
28 |
+
"vocab_size": 32768,
|
29 |
+
"quantization_config": {
|
30 |
+
"quant_method": "exl2",
|
31 |
+
"version": "0.0.18",
|
32 |
+
"bits": 3.75,
|
33 |
+
"head_bits": 6,
|
34 |
+
"calibration": {
|
35 |
+
"rows": 100,
|
36 |
+
"length": 2048,
|
37 |
+
"dataset": "(default)"
|
38 |
+
}
|
39 |
+
}
|
40 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 1,
|
4 |
+
"eos_token_id": 2,
|
5 |
+
"transformers_version": "4.34.0.dev0"
|
6 |
+
}
|
gitattributes
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
11 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
12 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
13 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
14 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
15 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
16 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
17 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
18 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
19 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
20 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
21 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
22 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
23 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
24 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
25 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
26 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
27 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
28 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
29 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
30 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
31 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
32 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
33 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
model.safetensors.index.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
output-00001-of-00008.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b1c2ff2ba5476b3df89330d08098251db24985d08020a613687664e20d9e792d
|
3 |
+
size 8581924752
|
output-00002-of-00008.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:161ec218c6a4e079570d9d0de990119f47d26860d816881e345185e09dda4365
|
3 |
+
size 8573208432
|
output-00003-of-00008.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8a256f8d323069917b9fbca66fb5ce6c3b414e8f05dc3ed4c45286e1af524c50
|
3 |
+
size 8558958648
|
output-00004-of-00008.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3c539862538fde337435a29f2954043b4bfe466b0453443543451f8adfa732fc
|
3 |
+
size 8553708680
|
output-00005-of-00008.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:18aeef37fc63451e856d24eb808801d4371361e3383ca769fa845f63bd93f252
|
3 |
+
size 8589855920
|
output-00006-of-00008.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bd50ac28545134deec452ec84e3234966380f12884c1108de902f1d6c79e1878
|
3 |
+
size 8573248536
|
output-00007-of-00008.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:74b999904f011d77e9097437a4e0eec2d2a1569fa8c6c3224dc06d1771f2d52c
|
3 |
+
size 8539662824
|
output-00008-of-00008.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:978d6eb3161eb053efecd9b4eff2dbec811239bca5ebbd81f3f8588d78b3d8a7
|
3 |
+
size 6296987240
|
special_tokens_map.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": "<s>",
|
3 |
+
"eos_token": "</s>",
|
4 |
+
"unk_token": "<unk>",
|
5 |
+
"b_inst": "[INST]",
|
6 |
+
"e_inst": "[/INST]"
|
7 |
+
}
|
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_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": false,
|
11 |
+
"special": true
|
12 |
+
},
|
13 |
+
"1": {
|
14 |
+
"content": "<s>",
|
15 |
+
"lstrip": false,
|
16 |
+
"normalized": true,
|
17 |
+
"rstrip": false,
|
18 |
+
"single_word": false,
|
19 |
+
"special": true
|
20 |
+
},
|
21 |
+
"2": {
|
22 |
+
"content": "</s>",
|
23 |
+
"lstrip": false,
|
24 |
+
"normalized": true,
|
25 |
+
"rstrip": false,
|
26 |
+
"single_word": false,
|
27 |
+
"special": true
|
28 |
+
}
|
29 |
+
},
|
30 |
+
"additional_special_tokens": [],
|
31 |
+
"bos_token": "<s>",
|
32 |
+
"chat_template": "{{bos_token}}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ ' [INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ ' ' + message['content'] + ' ' + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
|
33 |
+
"clean_up_tokenization_spaces": false,
|
34 |
+
"eos_token": "</s>",
|
35 |
+
"legacy": true,
|
36 |
+
"model_max_length": 1000000000000000019884624838656,
|
37 |
+
"pad_token": null,
|
38 |
+
"sp_model_kwargs": {},
|
39 |
+
"spaces_between_special_tokens": false,
|
40 |
+
"tokenizer_class": "LlamaTokenizer",
|
41 |
+
"unk_token": "<unk>",
|
42 |
+
"use_default_system_prompt": false
|
43 |
+
}
|