Upload folder using huggingface_hub
Browse files- README.md +1 -1
- config.json +1 -7
- generation_config.json +2 -5
- special_tokens_map.json +0 -0
- tokenizer_config.json +3 -7
README.md
CHANGED
@@ -28,7 +28,7 @@ license: apache-2.0
|
|
28 |
library_name: vllm
|
29 |
inference: false
|
30 |
base_model:
|
31 |
-
- mistralai/Mistral-Small-3.
|
32 |
extra_gated_description: >-
|
33 |
If you want to learn more about how we process your personal data, please read
|
34 |
our <a href="https://mistral.ai/terms/">Privacy Policy</a>.
|
|
|
28 |
library_name: vllm
|
29 |
inference: false
|
30 |
base_model:
|
31 |
+
- mistralai/Mistral-Small-3.1-24B-Base-2503
|
32 |
extra_gated_description: >-
|
33 |
If you want to learn more about how we process your personal data, please read
|
34 |
our <a href="https://mistral.ai/terms/">Privacy Policy</a>.
|
config.json
CHANGED
@@ -2,12 +2,9 @@
|
|
2 |
"architectures": [
|
3 |
"Mistral3ForConditionalGeneration"
|
4 |
],
|
5 |
-
"bos_token_id": 1,
|
6 |
-
"eos_token_id": 2,
|
7 |
"image_token_index": 10,
|
8 |
"model_type": "mistral3",
|
9 |
"multimodal_projector_bias": false,
|
10 |
-
"pad_token_id": 11,
|
11 |
"projector_hidden_act": "gelu",
|
12 |
"spatial_merge_size": 2,
|
13 |
"text_config": {
|
@@ -25,13 +22,11 @@
|
|
25 |
"rms_norm_eps": 1e-05,
|
26 |
"rope_theta": 1000000000.0,
|
27 |
"sliding_window": null,
|
28 |
-
"torch_dtype": "bfloat16",
|
29 |
"use_cache": true,
|
30 |
"vocab_size": 131072
|
31 |
},
|
32 |
"torch_dtype": "bfloat16",
|
33 |
"transformers_version": "4.52.4",
|
34 |
-
"unsloth_fixed": true,
|
35 |
"vision_config": {
|
36 |
"attention_dropout": 0.0,
|
37 |
"head_dim": 64,
|
@@ -45,8 +40,7 @@
|
|
45 |
"num_channels": 3,
|
46 |
"num_hidden_layers": 24,
|
47 |
"patch_size": 14,
|
48 |
-
"rope_theta": 10000.0
|
49 |
-
"torch_dtype": "bfloat16"
|
50 |
},
|
51 |
"vision_feature_layer": -1
|
52 |
}
|
|
|
2 |
"architectures": [
|
3 |
"Mistral3ForConditionalGeneration"
|
4 |
],
|
|
|
|
|
5 |
"image_token_index": 10,
|
6 |
"model_type": "mistral3",
|
7 |
"multimodal_projector_bias": false,
|
|
|
8 |
"projector_hidden_act": "gelu",
|
9 |
"spatial_merge_size": 2,
|
10 |
"text_config": {
|
|
|
22 |
"rms_norm_eps": 1e-05,
|
23 |
"rope_theta": 1000000000.0,
|
24 |
"sliding_window": null,
|
|
|
25 |
"use_cache": true,
|
26 |
"vocab_size": 131072
|
27 |
},
|
28 |
"torch_dtype": "bfloat16",
|
29 |
"transformers_version": "4.52.4",
|
|
|
30 |
"vision_config": {
|
31 |
"attention_dropout": 0.0,
|
32 |
"head_dim": 64,
|
|
|
40 |
"num_channels": 3,
|
41 |
"num_hidden_layers": 24,
|
42 |
"patch_size": 14,
|
43 |
+
"rope_theta": 10000.0
|
|
|
44 |
},
|
45 |
"vision_feature_layer": -1
|
46 |
}
|
generation_config.json
CHANGED
@@ -1,10 +1,7 @@
|
|
1 |
{
|
|
|
2 |
"bos_token_id": 1,
|
3 |
"eos_token_id": 2,
|
4 |
-
"
|
5 |
-
"do_sample": true,
|
6 |
-
"max_length": 131072,
|
7 |
"temperature": 0.15,
|
8 |
-
"top_p": 1.00,
|
9 |
-
"transformers_version": "4.52.4"
|
10 |
}
|
|
|
1 |
{
|
2 |
+
"_from_model_config": true,
|
3 |
"bos_token_id": 1,
|
4 |
"eos_token_id": 2,
|
5 |
+
"transformers_version": "4.52.4",
|
|
|
|
|
6 |
"temperature": 0.15,
|
|
|
|
|
7 |
}
|
special_tokens_map.json
CHANGED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
CHANGED
@@ -9011,12 +9011,8 @@
|
|
9011 |
"eos_token": "</s>",
|
9012 |
"extra_special_tokens": {},
|
9013 |
"legacy": true,
|
9014 |
-
"model_max_length":
|
9015 |
-
"pad_token": "<pad>",
|
9016 |
-
"padding_side": "left",
|
9017 |
-
"processor_class": "PixtralProcessor",
|
9018 |
"tokenizer_class": "LlamaTokenizerFast",
|
9019 |
"unk_token": "<unk>",
|
9020 |
-
"use_default_system_prompt": false
|
9021 |
-
|
9022 |
-
}
|
|
|
9011 |
"eos_token": "</s>",
|
9012 |
"extra_special_tokens": {},
|
9013 |
"legacy": true,
|
9014 |
+
"model_max_length": 1000000000000000019884624838656,
|
|
|
|
|
|
|
9015 |
"tokenizer_class": "LlamaTokenizerFast",
|
9016 |
"unk_token": "<unk>",
|
9017 |
+
"use_default_system_prompt": false
|
9018 |
+
}
|
|