Commit
•
c863612
1
Parent(s):
a14c9d5
Upload config.json with huggingface_hub
Browse files- config.json +60 -0
config.json
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "TIGER-Lab/Mantis-8B-siglip-llama3-pretraind",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlavaForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"ignore_index": -100,
|
| 7 |
+
"image_token_index": 128256,
|
| 8 |
+
"model_type": "llava",
|
| 9 |
+
"pad_token_id": 128257,
|
| 10 |
+
"projector_hidden_act": "gelu",
|
| 11 |
+
"quantization_config": {
|
| 12 |
+
"_load_in_4bit": true,
|
| 13 |
+
"_load_in_8bit": false,
|
| 14 |
+
"bnb_4bit_compute_dtype": "bfloat16",
|
| 15 |
+
"bnb_4bit_quant_storage": "bfloat16",
|
| 16 |
+
"bnb_4bit_quant_type": "nf4",
|
| 17 |
+
"bnb_4bit_use_double_quant": true,
|
| 18 |
+
"llm_int8_enable_fp32_cpu_offload": false,
|
| 19 |
+
"llm_int8_has_fp16_weight": false,
|
| 20 |
+
"llm_int8_skip_modules": [
|
| 21 |
+
"vision_tower"
|
| 22 |
+
],
|
| 23 |
+
"llm_int8_threshold": 6.0,
|
| 24 |
+
"load_in_4bit": true,
|
| 25 |
+
"load_in_8bit": false,
|
| 26 |
+
"quant_method": "bitsandbytes"
|
| 27 |
+
},
|
| 28 |
+
"text_config": {
|
| 29 |
+
"_name_or_path": "meta-llama/Meta-Llama-3-8B-Instruct",
|
| 30 |
+
"architectures": [
|
| 31 |
+
"LlamaForCausalLM"
|
| 32 |
+
],
|
| 33 |
+
"bos_token_id": 128000,
|
| 34 |
+
"eos_token_id": 128001,
|
| 35 |
+
"intermediate_size": 14336,
|
| 36 |
+
"max_position_embeddings": 8192,
|
| 37 |
+
"model_type": "llama",
|
| 38 |
+
"num_key_value_heads": 8,
|
| 39 |
+
"rms_norm_eps": 1e-05,
|
| 40 |
+
"rope_theta": 500000.0,
|
| 41 |
+
"torch_dtype": "bfloat16",
|
| 42 |
+
"vocab_size": 128258
|
| 43 |
+
},
|
| 44 |
+
"torch_dtype": "bfloat16",
|
| 45 |
+
"transformers_version": "4.41.0",
|
| 46 |
+
"vision_config": {
|
| 47 |
+
"hidden_act": "gelu_pytorch_tanh",
|
| 48 |
+
"hidden_size": 1152,
|
| 49 |
+
"image_size": 384,
|
| 50 |
+
"intermediate_size": 4304,
|
| 51 |
+
"layer_norm_eps": 1e-06,
|
| 52 |
+
"model_type": "siglip_vision_model",
|
| 53 |
+
"num_attention_heads": 16,
|
| 54 |
+
"num_hidden_layers": 27,
|
| 55 |
+
"patch_size": 14
|
| 56 |
+
},
|
| 57 |
+
"vision_feature_layer": -2,
|
| 58 |
+
"vision_feature_select_strategy": "default",
|
| 59 |
+
"vocab_size": 128258
|
| 60 |
+
}
|