Upload ./config.json with huggingface_hub
Browse files- config.json +46 -0
config.json
ADDED
|
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"X": [
|
| 3 |
+
"Video",
|
| 4 |
+
"Image"
|
| 5 |
+
],
|
| 6 |
+
"_name_or_path": "lmsys/vicuna-7b-v1.5",
|
| 7 |
+
"architectures": [
|
| 8 |
+
"LlavaLlamaForCausalLM"
|
| 9 |
+
],
|
| 10 |
+
"attention_bias": false,
|
| 11 |
+
"attention_dropout": 0.0,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"eos_token_id": 2,
|
| 14 |
+
"freeze_mm_mlp_adapter": false,
|
| 15 |
+
"hidden_act": "silu",
|
| 16 |
+
"hidden_size": 4096,
|
| 17 |
+
"image_aspect_ratio": "pad",
|
| 18 |
+
"image_grid_pinpoints": null,
|
| 19 |
+
"initializer_range": 0.02,
|
| 20 |
+
"intermediate_size": 11008,
|
| 21 |
+
"max_position_embeddings": 4096,
|
| 22 |
+
"mm_hidden_size": 1024,
|
| 23 |
+
"mm_image_tower": "LanguageBind/LanguageBind_Image",
|
| 24 |
+
"mm_projector_type": "mlp2x_gelu",
|
| 25 |
+
"mm_use_x_patch_token": false,
|
| 26 |
+
"mm_use_x_start_end": false,
|
| 27 |
+
"mm_video_tower": "LanguageBind/LanguageBind_Video_merge",
|
| 28 |
+
"mm_vision_select_feature": "patch",
|
| 29 |
+
"mm_vision_select_layer": -2,
|
| 30 |
+
"model_type": "llava_llama",
|
| 31 |
+
"num_attention_heads": 32,
|
| 32 |
+
"num_hidden_layers": 32,
|
| 33 |
+
"num_key_value_heads": 32,
|
| 34 |
+
"pad_token_id": 0,
|
| 35 |
+
"pretraining_tp": 1,
|
| 36 |
+
"rms_norm_eps": 1e-05,
|
| 37 |
+
"rope_scaling": null,
|
| 38 |
+
"rope_theta": 10000.0,
|
| 39 |
+
"tie_word_embeddings": false,
|
| 40 |
+
"torch_dtype": "bfloat16",
|
| 41 |
+
"transformers_version": "4.36.2",
|
| 42 |
+
"tune_mm_mlp_adapter": false,
|
| 43 |
+
"use_cache": true,
|
| 44 |
+
"use_mm_proj": true,
|
| 45 |
+
"vocab_size": 32000
|
| 46 |
+
}
|