457999f33ec35d0d01ef0e34ad85c32ee35dd566ebb1326528e294cc97125799
Browse files- 7b/config.json +51 -0
- 7b/mm_projector.bin +3 -0
- 7b/trainer_state.json +0 -0
7b/config.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "Qwen/Qwen2-7B-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": 3584,
|
| 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": 18944,
|
| 17 |
+
"max_position_embeddings": 32768,
|
| 18 |
+
"max_window_layers": 28,
|
| 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": 28,
|
| 33 |
+
"num_hidden_layers": 28,
|
| 34 |
+
"num_key_value_heads": 4,
|
| 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 |
+
}
|
7b/mm_projector.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:08fd70becae179cfc7942bd0759c5ef179b71b9459e740f65c6f758d789a6f1e
|
| 3 |
+
size 33964088
|
7b/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|