Upload folder using huggingface_hub (#3)
Browse files- c31805c23253a58aff661031695d19b9d2c0bf0da7ca92645a6e9ee383d0de50 (e429c10c74b538c505a5e451f642dad756c3d13f)
- 72b/config.json +51 -0
- 72b/mm_projector.bin +3 -0
- 72b/trainer_state.json +0 -0
72b/config.json
ADDED
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "Qwen/Qwen2-72B-Instruct",
|
3 |
+
"architectures": [
|
4 |
+
"Qwen2ForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_dropout": 0.0,
|
7 |
+
"bos_token_id": 151643,
|
8 |
+
"eos_token_id": 151645,
|
9 |
+
"hidden_act": "silu",
|
10 |
+
"hidden_size": 8192,
|
11 |
+
"image_aspect_ratio": "square",
|
12 |
+
"image_crop_resolution": null,
|
13 |
+
"image_grid_pinpoints": null,
|
14 |
+
"image_split_resolution": null,
|
15 |
+
"initializer_range": 0.02,
|
16 |
+
"intermediate_size": 29568,
|
17 |
+
"max_position_embeddings": 32768,
|
18 |
+
"max_window_layers": 80,
|
19 |
+
"mm_hidden_size": 1152,
|
20 |
+
"mm_patch_merge_type": "flat",
|
21 |
+
"mm_projector_lr": null,
|
22 |
+
"mm_projector_type": "mlp2x_gelu",
|
23 |
+
"mm_resampler_type": null,
|
24 |
+
"mm_tunable_parts": "mm_mlp_adapter",
|
25 |
+
"mm_use_im_patch_token": false,
|
26 |
+
"mm_use_im_start_end": false,
|
27 |
+
"mm_vision_select_feature": "patch",
|
28 |
+
"mm_vision_select_layer": -2,
|
29 |
+
"mm_vision_tower": "google/siglip-so400m-patch14-384",
|
30 |
+
"mm_vision_tower_lr": null,
|
31 |
+
"model_type": "qwen2",
|
32 |
+
"num_attention_heads": 64,
|
33 |
+
"num_hidden_layers": 80,
|
34 |
+
"num_key_value_heads": 8,
|
35 |
+
"pos_skipping_range": 4096,
|
36 |
+
"rms_norm_eps": 1e-06,
|
37 |
+
"rope_scaling": null,
|
38 |
+
"rope_theta": 1000000.0,
|
39 |
+
"sliding_window": 131072,
|
40 |
+
"tie_word_embeddings": false,
|
41 |
+
"tokenizer_model_max_length": 8192,
|
42 |
+
"tokenizer_padding_side": "right",
|
43 |
+
"torch_dtype": "bfloat16",
|
44 |
+
"transformers_version": "4.40.0.dev0",
|
45 |
+
"use_cache": true,
|
46 |
+
"use_mm_proj": true,
|
47 |
+
"use_pos_skipping": false,
|
48 |
+
"use_sliding_window": false,
|
49 |
+
"vision_tower_pretrained": null,
|
50 |
+
"vocab_size": 152064
|
51 |
+
}
|
72b/mm_projector.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:20551dfa7e58956bdf8a505bae4788aa8a4184a9c9ca175a30cb7a52c567c5f4
|
3 |
+
size 153126968
|
72b/trainer_state.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|