2.0.0 preview
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +23 -0
- devkit/added_tokens.json +4 -0
- devkit/chat_template.jinja +2 -0
- devkit/chat_template.json +3 -0
- devkit/config.json +38 -0
- devkit/generation_config.json +7 -0
- devkit/models--llava-hf--llava-1.5-7b-hf_language_embeddings.npy +2 -2
- devkit/precision.json +832 -0
- devkit/preprocessor_config.json +28 -0
- devkit/processor_config.json +7 -0
- devkit/special_tokens_map.json +30 -0
- devkit/tokenizer.json +0 -0
- devkit/tokenizer_config.json +64 -0
- devkit/vlm_config.json +24 -12
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token0_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1035_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1135_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1235_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1335_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token135_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1435_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1535_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1635_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1735_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1835_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1935_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token2035_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token235_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token335_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token35_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token435_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token535_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token635_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token735_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token835_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token935_stage1_mla.elf +3 -0
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer0_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer10_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer11_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer12_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer13_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer14_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer15_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer16_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer17_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer18_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer19_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer1_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer20_stage1_mla.elf +2 -2
- elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer21_stage1_mla.elf +2 -2
.gitattributes
CHANGED
|
@@ -176,3 +176,26 @@ elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_pre_layer7_stage1_mla.el
|
|
| 176 |
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_pre_layer8_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 177 |
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_pre_layer9_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 178 |
elf_files/models--llava-hf--llava-1.5-7b-hf_vision_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 176 |
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_pre_layer8_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 177 |
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_pre_layer9_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 178 |
elf_files/models--llava-hf--llava-1.5-7b-hf_vision_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 179 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1035_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 180 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1135_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 181 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1235_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 182 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1335_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 183 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1435_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 184 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1535_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 185 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1635_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 186 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1735_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 187 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1835_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 188 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1935_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 189 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token2035_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 190 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token635_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 191 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token735_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 192 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token835_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 193 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token935_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 194 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_cache_token1151_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 195 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_cache_token1279_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 196 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_cache_token1407_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 197 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_cache_token1535_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 198 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_cache_token1663_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 199 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_cache_token1791_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 200 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_cache_token1919_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
| 201 |
+
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n1_cache_token2047_stage1_mla.elf filter=lfs diff=lfs merge=lfs -text
|
devkit/added_tokens.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<image>": 32000,
|
| 3 |
+
"<pad>": 32001
|
| 4 |
+
}
|
devkit/chat_template.jinja
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{% for message in messages %}{% if message['role'] != 'system' %}{{ message['role'].upper() + ': '}}{% endif %}{# Render all images first #}{% for content in message['content'] | selectattr('type', 'equalto', 'image') %}{{ '<image>
|
| 2 |
+
' }}{% endfor %}{# Render all text next #}{% if message['role'] != 'assistant' %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{{ content['text'] + ' '}}{% endfor %}{% else %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{% generation %}{{ content['text'] + ' '}}{% endgeneration %}{% endfor %}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ 'ASSISTANT:' }}{% endif %}
|
devkit/chat_template.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chat_template": "{% for message in messages %}{% if message['role'] != 'system' %}{{ message['role'].upper() + ': '}}{% endif %}{# Render all images first #}{% for content in message['content'] | selectattr('type', 'equalto', 'image') %}{{ '<image>\n' }}{% endfor %}{# Render all text next #}{% if message['role'] != 'assistant' %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{{ content['text'] + ' '}}{% endfor %}{% else %}{% for content in message['content'] | selectattr('type', 'equalto', 'text') %}{% generation %}{{ content['text'] + ' '}}{% endgeneration %}{% endfor %}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ 'ASSISTANT:' }}{% endif %}"
|
| 3 |
+
}
|
devkit/config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"LlavaForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"ignore_index": -100,
|
| 6 |
+
"image_token_index": 32000,
|
| 7 |
+
"model_type": "llava",
|
| 8 |
+
"pad_token_id": 32001,
|
| 9 |
+
"projector_hidden_act": "gelu",
|
| 10 |
+
"text_config": {
|
| 11 |
+
"_name_or_path": "lmsys/vicuna-7b-v1.5",
|
| 12 |
+
"architectures": [
|
| 13 |
+
"LlamaForCausalLM"
|
| 14 |
+
],
|
| 15 |
+
"max_position_embeddings": 4096,
|
| 16 |
+
"model_type": "llama",
|
| 17 |
+
"rms_norm_eps": 1e-05,
|
| 18 |
+
"torch_dtype": "float16",
|
| 19 |
+
"vocab_size": 32064
|
| 20 |
+
},
|
| 21 |
+
"tie_word_embeddings": false,
|
| 22 |
+
"torch_dtype": "float16",
|
| 23 |
+
"transformers_version": "4.36.0.dev0",
|
| 24 |
+
"vision_config": {
|
| 25 |
+
"hidden_size": 1024,
|
| 26 |
+
"image_size": 336,
|
| 27 |
+
"intermediate_size": 4096,
|
| 28 |
+
"model_type": "clip_vision_model",
|
| 29 |
+
"num_attention_heads": 16,
|
| 30 |
+
"num_hidden_layers": 24,
|
| 31 |
+
"patch_size": 14,
|
| 32 |
+
"projection_dim": 768,
|
| 33 |
+
"vocab_size": 32000
|
| 34 |
+
},
|
| 35 |
+
"vision_feature_layer": -2,
|
| 36 |
+
"vision_feature_select_strategy": "default",
|
| 37 |
+
"vocab_size": 32064
|
| 38 |
+
}
|
devkit/generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 32001,
|
| 6 |
+
"transformers_version": "4.36.0.dev0"
|
| 7 |
+
}
|
devkit/models--llava-hf--llava-1.5-7b-hf_language_embeddings.npy
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1158e5dbf1235817e96fdeeb742c74d2b0d8a5530544ea39afe200f36b060361
|
| 3 |
+
size 262668416
|
devkit/precision.json
ADDED
|
@@ -0,0 +1,832 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"part": "group_pre",
|
| 4 |
+
"idx": 0,
|
| 5 |
+
"precision": "A_BF16_W_INT4"
|
| 6 |
+
},
|
| 7 |
+
{
|
| 8 |
+
"part": "group_pre",
|
| 9 |
+
"idx": 1,
|
| 10 |
+
"precision": "A_BF16_W_INT4"
|
| 11 |
+
},
|
| 12 |
+
{
|
| 13 |
+
"part": "group_pre",
|
| 14 |
+
"idx": 2,
|
| 15 |
+
"precision": "A_BF16_W_INT4"
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"part": "group_pre",
|
| 19 |
+
"idx": 3,
|
| 20 |
+
"precision": "A_BF16_W_INT4"
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"part": "group_pre",
|
| 24 |
+
"idx": 4,
|
| 25 |
+
"precision": "A_BF16_W_INT4"
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"part": "group_pre",
|
| 29 |
+
"idx": 5,
|
| 30 |
+
"precision": "A_BF16_W_INT4"
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"part": "group_pre",
|
| 34 |
+
"idx": 6,
|
| 35 |
+
"precision": "A_BF16_W_INT4"
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"part": "group_pre",
|
| 39 |
+
"idx": 7,
|
| 40 |
+
"precision": "A_BF16_W_INT4"
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"part": "group_pre",
|
| 44 |
+
"idx": 8,
|
| 45 |
+
"precision": "A_BF16_W_INT4"
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"part": "group_pre",
|
| 49 |
+
"idx": 9,
|
| 50 |
+
"precision": "A_BF16_W_INT4"
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"part": "group_pre",
|
| 54 |
+
"idx": 10,
|
| 55 |
+
"precision": "A_BF16_W_INT4"
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"part": "group_pre",
|
| 59 |
+
"idx": 11,
|
| 60 |
+
"precision": "A_BF16_W_INT4"
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"part": "group_pre",
|
| 64 |
+
"idx": 12,
|
| 65 |
+
"precision": "A_BF16_W_INT4"
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"part": "group_pre",
|
| 69 |
+
"idx": 13,
|
| 70 |
+
"precision": "A_BF16_W_INT4"
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"part": "group_pre",
|
| 74 |
+
"idx": 14,
|
| 75 |
+
"precision": "A_BF16_W_INT4"
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"part": "group_pre",
|
| 79 |
+
"idx": 15,
|
| 80 |
+
"precision": "A_BF16_W_INT4"
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"part": "group_pre",
|
| 84 |
+
"idx": 16,
|
| 85 |
+
"precision": "A_BF16_W_INT4"
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"part": "group_pre",
|
| 89 |
+
"idx": 17,
|
| 90 |
+
"precision": "A_BF16_W_INT4"
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"part": "group_pre",
|
| 94 |
+
"idx": 18,
|
| 95 |
+
"precision": "A_BF16_W_INT4"
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"part": "group_pre",
|
| 99 |
+
"idx": 19,
|
| 100 |
+
"precision": "A_BF16_W_INT4"
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"part": "group_pre",
|
| 104 |
+
"idx": 20,
|
| 105 |
+
"precision": "A_BF16_W_INT4"
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"part": "group_pre",
|
| 109 |
+
"idx": 21,
|
| 110 |
+
"precision": "A_BF16_W_INT4"
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"part": "group_pre",
|
| 114 |
+
"idx": 22,
|
| 115 |
+
"precision": "A_BF16_W_INT4"
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"part": "group_pre",
|
| 119 |
+
"idx": 23,
|
| 120 |
+
"precision": "A_BF16_W_INT4"
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"part": "group_pre",
|
| 124 |
+
"idx": 24,
|
| 125 |
+
"precision": "A_BF16_W_INT4"
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"part": "group_pre",
|
| 129 |
+
"idx": 25,
|
| 130 |
+
"precision": "A_BF16_W_INT4"
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"part": "group_pre",
|
| 134 |
+
"idx": 26,
|
| 135 |
+
"precision": "A_BF16_W_INT4"
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"part": "group_pre",
|
| 139 |
+
"idx": 27,
|
| 140 |
+
"precision": "A_BF16_W_INT4"
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"part": "group_pre",
|
| 144 |
+
"idx": 28,
|
| 145 |
+
"precision": "A_BF16_W_INT4"
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"part": "group_pre",
|
| 149 |
+
"idx": 29,
|
| 150 |
+
"precision": "A_BF16_W_INT4"
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"part": "group_pre",
|
| 154 |
+
"idx": 30,
|
| 155 |
+
"precision": "A_BF16_W_INT4"
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"part": "group_pre",
|
| 159 |
+
"idx": 31,
|
| 160 |
+
"precision": "A_BF16_W_INT4"
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"part": "group_post",
|
| 164 |
+
"idx": 0,
|
| 165 |
+
"precision": "A_BF16_W_INT4"
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"part": "group_post",
|
| 169 |
+
"idx": 1,
|
| 170 |
+
"precision": "A_BF16_W_INT4"
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"part": "group_post",
|
| 174 |
+
"idx": 2,
|
| 175 |
+
"precision": "A_BF16_W_INT4"
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"part": "group_post",
|
| 179 |
+
"idx": 3,
|
| 180 |
+
"precision": "A_BF16_W_INT4"
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"part": "group_post",
|
| 184 |
+
"idx": 4,
|
| 185 |
+
"precision": "A_BF16_W_INT4"
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"part": "group_post",
|
| 189 |
+
"idx": 5,
|
| 190 |
+
"precision": "A_BF16_W_INT4"
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"part": "group_post",
|
| 194 |
+
"idx": 6,
|
| 195 |
+
"precision": "A_BF16_W_INT4"
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"part": "group_post",
|
| 199 |
+
"idx": 7,
|
| 200 |
+
"precision": "A_BF16_W_INT4"
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"part": "group_post",
|
| 204 |
+
"idx": 8,
|
| 205 |
+
"precision": "A_BF16_W_INT4"
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"part": "group_post",
|
| 209 |
+
"idx": 9,
|
| 210 |
+
"precision": "A_BF16_W_INT4"
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"part": "group_post",
|
| 214 |
+
"idx": 10,
|
| 215 |
+
"precision": "A_BF16_W_INT4"
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"part": "group_post",
|
| 219 |
+
"idx": 11,
|
| 220 |
+
"precision": "A_BF16_W_INT4"
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"part": "group_post",
|
| 224 |
+
"idx": 12,
|
| 225 |
+
"precision": "A_BF16_W_INT4"
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"part": "group_post",
|
| 229 |
+
"idx": 13,
|
| 230 |
+
"precision": "A_BF16_W_INT4"
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"part": "group_post",
|
| 234 |
+
"idx": 14,
|
| 235 |
+
"precision": "A_BF16_W_INT4"
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"part": "group_post",
|
| 239 |
+
"idx": 15,
|
| 240 |
+
"precision": "A_BF16_W_INT4"
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"part": "group_post",
|
| 244 |
+
"idx": 16,
|
| 245 |
+
"precision": "A_BF16_W_INT4"
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"part": "group_post",
|
| 249 |
+
"idx": 17,
|
| 250 |
+
"precision": "A_BF16_W_INT4"
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"part": "group_post",
|
| 254 |
+
"idx": 18,
|
| 255 |
+
"precision": "A_BF16_W_INT4"
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"part": "group_post",
|
| 259 |
+
"idx": 19,
|
| 260 |
+
"precision": "A_BF16_W_INT4"
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"part": "group_post",
|
| 264 |
+
"idx": 20,
|
| 265 |
+
"precision": "A_BF16_W_INT4"
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"part": "group_post",
|
| 269 |
+
"idx": 21,
|
| 270 |
+
"precision": "A_BF16_W_INT4"
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"part": "group_post",
|
| 274 |
+
"idx": 22,
|
| 275 |
+
"precision": "A_BF16_W_INT4"
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"part": "group_post",
|
| 279 |
+
"idx": 23,
|
| 280 |
+
"precision": "A_BF16_W_INT4"
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"part": "group_post",
|
| 284 |
+
"idx": 24,
|
| 285 |
+
"precision": "A_BF16_W_INT4"
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"part": "group_post",
|
| 289 |
+
"idx": 25,
|
| 290 |
+
"precision": "A_BF16_W_INT4"
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"part": "group_post",
|
| 294 |
+
"idx": 26,
|
| 295 |
+
"precision": "A_BF16_W_INT4"
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"part": "group_post",
|
| 299 |
+
"idx": 27,
|
| 300 |
+
"precision": "A_BF16_W_INT4"
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"part": "group_post",
|
| 304 |
+
"idx": 28,
|
| 305 |
+
"precision": "A_BF16_W_INT4"
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"part": "group_post",
|
| 309 |
+
"idx": 29,
|
| 310 |
+
"precision": "A_BF16_W_INT4"
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"part": "group_post",
|
| 314 |
+
"idx": 30,
|
| 315 |
+
"precision": "A_BF16_W_INT4"
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"part": "group_cache",
|
| 319 |
+
"idx": 0,
|
| 320 |
+
"precision": "A_BF16_W_INT4"
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"part": "group_cache",
|
| 324 |
+
"idx": 35,
|
| 325 |
+
"precision": "A_BF16_W_INT4"
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"part": "group_cache",
|
| 329 |
+
"idx": 135,
|
| 330 |
+
"precision": "A_BF16_W_INT4"
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"part": "group_cache",
|
| 334 |
+
"idx": 235,
|
| 335 |
+
"precision": "A_BF16_W_INT4"
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"part": "group_cache",
|
| 339 |
+
"idx": 335,
|
| 340 |
+
"precision": "A_BF16_W_INT4"
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"part": "group_cache",
|
| 344 |
+
"idx": 435,
|
| 345 |
+
"precision": "A_BF16_W_INT4"
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"part": "group_cache",
|
| 349 |
+
"idx": 535,
|
| 350 |
+
"precision": "A_BF16_W_INT4"
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"part": "group_cache",
|
| 354 |
+
"idx": 635,
|
| 355 |
+
"precision": "A_BF16_W_INT4"
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"part": "group_cache",
|
| 359 |
+
"idx": 735,
|
| 360 |
+
"precision": "A_BF16_W_INT4"
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"part": "group_cache",
|
| 364 |
+
"idx": 835,
|
| 365 |
+
"precision": "A_BF16_W_INT4"
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"part": "group_cache",
|
| 369 |
+
"idx": 935,
|
| 370 |
+
"precision": "A_BF16_W_INT4"
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"part": "group_cache",
|
| 374 |
+
"idx": 1035,
|
| 375 |
+
"precision": "A_BF16_W_INT4"
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"part": "group_cache",
|
| 379 |
+
"idx": 1135,
|
| 380 |
+
"precision": "A_BF16_W_INT4"
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"part": "group_cache",
|
| 384 |
+
"idx": 1235,
|
| 385 |
+
"precision": "A_BF16_W_INT4"
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"part": "group_cache",
|
| 389 |
+
"idx": 1335,
|
| 390 |
+
"precision": "A_BF16_W_INT4"
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"part": "group_cache",
|
| 394 |
+
"idx": 1435,
|
| 395 |
+
"precision": "A_BF16_W_INT4"
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"part": "group_cache",
|
| 399 |
+
"idx": 1535,
|
| 400 |
+
"precision": "A_BF16_W_INT4"
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"part": "group_cache",
|
| 404 |
+
"idx": 1635,
|
| 405 |
+
"precision": "A_BF16_W_INT4"
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"part": "group_cache",
|
| 409 |
+
"idx": 1735,
|
| 410 |
+
"precision": "A_BF16_W_INT4"
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"part": "group_cache",
|
| 414 |
+
"idx": 1835,
|
| 415 |
+
"precision": "A_BF16_W_INT4"
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"part": "group_cache",
|
| 419 |
+
"idx": 1935,
|
| 420 |
+
"precision": "A_BF16_W_INT4"
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"part": "group_cache",
|
| 424 |
+
"idx": 2035,
|
| 425 |
+
"precision": "A_BF16_W_INT4"
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"part": "single_pre",
|
| 429 |
+
"idx": 0,
|
| 430 |
+
"precision": "A_BF16_W_INT4"
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"part": "single_pre",
|
| 434 |
+
"idx": 1,
|
| 435 |
+
"precision": "A_BF16_W_INT4"
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"part": "single_pre",
|
| 439 |
+
"idx": 2,
|
| 440 |
+
"precision": "A_BF16_W_INT4"
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"part": "single_pre",
|
| 444 |
+
"idx": 3,
|
| 445 |
+
"precision": "A_BF16_W_INT4"
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"part": "single_pre",
|
| 449 |
+
"idx": 4,
|
| 450 |
+
"precision": "A_BF16_W_INT4"
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"part": "single_pre",
|
| 454 |
+
"idx": 5,
|
| 455 |
+
"precision": "A_BF16_W_INT4"
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"part": "single_pre",
|
| 459 |
+
"idx": 6,
|
| 460 |
+
"precision": "A_BF16_W_INT4"
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"part": "single_pre",
|
| 464 |
+
"idx": 7,
|
| 465 |
+
"precision": "A_BF16_W_INT4"
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"part": "single_pre",
|
| 469 |
+
"idx": 8,
|
| 470 |
+
"precision": "A_BF16_W_INT4"
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"part": "single_pre",
|
| 474 |
+
"idx": 9,
|
| 475 |
+
"precision": "A_BF16_W_INT4"
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"part": "single_pre",
|
| 479 |
+
"idx": 10,
|
| 480 |
+
"precision": "A_BF16_W_INT4"
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"part": "single_pre",
|
| 484 |
+
"idx": 11,
|
| 485 |
+
"precision": "A_BF16_W_INT4"
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"part": "single_pre",
|
| 489 |
+
"idx": 12,
|
| 490 |
+
"precision": "A_BF16_W_INT4"
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"part": "single_pre",
|
| 494 |
+
"idx": 13,
|
| 495 |
+
"precision": "A_BF16_W_INT4"
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"part": "single_pre",
|
| 499 |
+
"idx": 14,
|
| 500 |
+
"precision": "A_BF16_W_INT4"
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"part": "single_pre",
|
| 504 |
+
"idx": 15,
|
| 505 |
+
"precision": "A_BF16_W_INT4"
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"part": "single_pre",
|
| 509 |
+
"idx": 16,
|
| 510 |
+
"precision": "A_BF16_W_INT4"
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"part": "single_pre",
|
| 514 |
+
"idx": 17,
|
| 515 |
+
"precision": "A_BF16_W_INT4"
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"part": "single_pre",
|
| 519 |
+
"idx": 18,
|
| 520 |
+
"precision": "A_BF16_W_INT4"
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"part": "single_pre",
|
| 524 |
+
"idx": 19,
|
| 525 |
+
"precision": "A_BF16_W_INT4"
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"part": "single_pre",
|
| 529 |
+
"idx": 20,
|
| 530 |
+
"precision": "A_BF16_W_INT4"
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"part": "single_pre",
|
| 534 |
+
"idx": 21,
|
| 535 |
+
"precision": "A_BF16_W_INT4"
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"part": "single_pre",
|
| 539 |
+
"idx": 22,
|
| 540 |
+
"precision": "A_BF16_W_INT4"
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"part": "single_pre",
|
| 544 |
+
"idx": 23,
|
| 545 |
+
"precision": "A_BF16_W_INT4"
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"part": "single_pre",
|
| 549 |
+
"idx": 24,
|
| 550 |
+
"precision": "A_BF16_W_INT4"
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"part": "single_pre",
|
| 554 |
+
"idx": 25,
|
| 555 |
+
"precision": "A_BF16_W_INT4"
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"part": "single_pre",
|
| 559 |
+
"idx": 26,
|
| 560 |
+
"precision": "A_BF16_W_INT4"
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"part": "single_pre",
|
| 564 |
+
"idx": 27,
|
| 565 |
+
"precision": "A_BF16_W_INT4"
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"part": "single_pre",
|
| 569 |
+
"idx": 28,
|
| 570 |
+
"precision": "A_BF16_W_INT4"
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"part": "single_pre",
|
| 574 |
+
"idx": 29,
|
| 575 |
+
"precision": "A_BF16_W_INT4"
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"part": "single_pre",
|
| 579 |
+
"idx": 30,
|
| 580 |
+
"precision": "A_BF16_W_INT4"
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"part": "single_pre",
|
| 584 |
+
"idx": 31,
|
| 585 |
+
"precision": "A_BF16_W_INT4"
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"part": "single_post",
|
| 589 |
+
"idx": 0,
|
| 590 |
+
"precision": "A_BF16_W_INT4"
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"part": "single_post",
|
| 594 |
+
"idx": 1,
|
| 595 |
+
"precision": "A_BF16_W_INT4"
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"part": "single_post",
|
| 599 |
+
"idx": 2,
|
| 600 |
+
"precision": "A_BF16_W_INT4"
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"part": "single_post",
|
| 604 |
+
"idx": 3,
|
| 605 |
+
"precision": "A_BF16_W_INT4"
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"part": "single_post",
|
| 609 |
+
"idx": 4,
|
| 610 |
+
"precision": "A_BF16_W_INT4"
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"part": "single_post",
|
| 614 |
+
"idx": 5,
|
| 615 |
+
"precision": "A_BF16_W_INT4"
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"part": "single_post",
|
| 619 |
+
"idx": 6,
|
| 620 |
+
"precision": "A_BF16_W_INT4"
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"part": "single_post",
|
| 624 |
+
"idx": 7,
|
| 625 |
+
"precision": "A_BF16_W_INT4"
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"part": "single_post",
|
| 629 |
+
"idx": 8,
|
| 630 |
+
"precision": "A_BF16_W_INT4"
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"part": "single_post",
|
| 634 |
+
"idx": 9,
|
| 635 |
+
"precision": "A_BF16_W_INT4"
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"part": "single_post",
|
| 639 |
+
"idx": 10,
|
| 640 |
+
"precision": "A_BF16_W_INT4"
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"part": "single_post",
|
| 644 |
+
"idx": 11,
|
| 645 |
+
"precision": "A_BF16_W_INT4"
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"part": "single_post",
|
| 649 |
+
"idx": 12,
|
| 650 |
+
"precision": "A_BF16_W_INT4"
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"part": "single_post",
|
| 654 |
+
"idx": 13,
|
| 655 |
+
"precision": "A_BF16_W_INT4"
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"part": "single_post",
|
| 659 |
+
"idx": 14,
|
| 660 |
+
"precision": "A_BF16_W_INT4"
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"part": "single_post",
|
| 664 |
+
"idx": 15,
|
| 665 |
+
"precision": "A_BF16_W_INT4"
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"part": "single_post",
|
| 669 |
+
"idx": 16,
|
| 670 |
+
"precision": "A_BF16_W_INT4"
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"part": "single_post",
|
| 674 |
+
"idx": 17,
|
| 675 |
+
"precision": "A_BF16_W_INT4"
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"part": "single_post",
|
| 679 |
+
"idx": 18,
|
| 680 |
+
"precision": "A_BF16_W_INT4"
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"part": "single_post",
|
| 684 |
+
"idx": 19,
|
| 685 |
+
"precision": "A_BF16_W_INT4"
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"part": "single_post",
|
| 689 |
+
"idx": 20,
|
| 690 |
+
"precision": "A_BF16_W_INT4"
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"part": "single_post",
|
| 694 |
+
"idx": 21,
|
| 695 |
+
"precision": "A_BF16_W_INT4"
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"part": "single_post",
|
| 699 |
+
"idx": 22,
|
| 700 |
+
"precision": "A_BF16_W_INT4"
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"part": "single_post",
|
| 704 |
+
"idx": 23,
|
| 705 |
+
"precision": "A_BF16_W_INT4"
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"part": "single_post",
|
| 709 |
+
"idx": 24,
|
| 710 |
+
"precision": "A_BF16_W_INT4"
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"part": "single_post",
|
| 714 |
+
"idx": 25,
|
| 715 |
+
"precision": "A_BF16_W_INT4"
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"part": "single_post",
|
| 719 |
+
"idx": 26,
|
| 720 |
+
"precision": "A_BF16_W_INT4"
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"part": "single_post",
|
| 724 |
+
"idx": 27,
|
| 725 |
+
"precision": "A_BF16_W_INT4"
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"part": "single_post",
|
| 729 |
+
"idx": 28,
|
| 730 |
+
"precision": "A_BF16_W_INT4"
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"part": "single_post",
|
| 734 |
+
"idx": 29,
|
| 735 |
+
"precision": "A_BF16_W_INT4"
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"part": "single_post",
|
| 739 |
+
"idx": 30,
|
| 740 |
+
"precision": "A_BF16_W_INT4"
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"part": "single_post",
|
| 744 |
+
"idx": 31,
|
| 745 |
+
"precision": "A_BF16_W_INT4"
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"part": "single_cache",
|
| 749 |
+
"idx": 127,
|
| 750 |
+
"precision": "A_BF16_W_INT4"
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"part": "single_cache",
|
| 754 |
+
"idx": 255,
|
| 755 |
+
"precision": "A_BF16_W_INT4"
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"part": "single_cache",
|
| 759 |
+
"idx": 383,
|
| 760 |
+
"precision": "A_BF16_W_INT4"
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"part": "single_cache",
|
| 764 |
+
"idx": 511,
|
| 765 |
+
"precision": "A_BF16_W_INT4"
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"part": "single_cache",
|
| 769 |
+
"idx": 639,
|
| 770 |
+
"precision": "A_BF16_W_INT4"
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"part": "single_cache",
|
| 774 |
+
"idx": 767,
|
| 775 |
+
"precision": "A_BF16_W_INT4"
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"part": "single_cache",
|
| 779 |
+
"idx": 895,
|
| 780 |
+
"precision": "A_BF16_W_INT4"
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"part": "single_cache",
|
| 784 |
+
"idx": 1023,
|
| 785 |
+
"precision": "A_BF16_W_INT4"
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"part": "single_cache",
|
| 789 |
+
"idx": 1151,
|
| 790 |
+
"precision": "A_BF16_W_INT4"
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"part": "single_cache",
|
| 794 |
+
"idx": 1279,
|
| 795 |
+
"precision": "A_BF16_W_INT4"
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"part": "single_cache",
|
| 799 |
+
"idx": 1407,
|
| 800 |
+
"precision": "A_BF16_W_INT4"
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"part": "single_cache",
|
| 804 |
+
"idx": 1535,
|
| 805 |
+
"precision": "A_BF16_W_INT4"
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"part": "single_cache",
|
| 809 |
+
"idx": 1663,
|
| 810 |
+
"precision": "A_BF16_W_INT4"
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"part": "single_cache",
|
| 814 |
+
"idx": 1791,
|
| 815 |
+
"precision": "A_BF16_W_INT4"
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"part": "single_cache",
|
| 819 |
+
"idx": 1919,
|
| 820 |
+
"precision": "A_BF16_W_INT4"
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"part": "single_cache",
|
| 824 |
+
"idx": 2047,
|
| 825 |
+
"precision": "A_BF16_W_INT4"
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"part": "vision",
|
| 829 |
+
"idx": 0,
|
| 830 |
+
"precision": "BF16"
|
| 831 |
+
}
|
| 832 |
+
]
|
devkit/preprocessor_config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"crop_size": {
|
| 3 |
+
"height": 336,
|
| 4 |
+
"width": 336
|
| 5 |
+
},
|
| 6 |
+
"do_center_crop": true,
|
| 7 |
+
"do_convert_rgb": true,
|
| 8 |
+
"do_normalize": true,
|
| 9 |
+
"do_rescale": true,
|
| 10 |
+
"do_resize": true,
|
| 11 |
+
"image_mean": [
|
| 12 |
+
0.48145466,
|
| 13 |
+
0.4578275,
|
| 14 |
+
0.40821073
|
| 15 |
+
],
|
| 16 |
+
"image_processor_type": "CLIPImageProcessor",
|
| 17 |
+
"image_std": [
|
| 18 |
+
0.26862954,
|
| 19 |
+
0.26130258,
|
| 20 |
+
0.27577711
|
| 21 |
+
],
|
| 22 |
+
"processor_class": "LlavaProcessor",
|
| 23 |
+
"resample": 3,
|
| 24 |
+
"rescale_factor": 0.00392156862745098,
|
| 25 |
+
"size": {
|
| 26 |
+
"shortest_edge": 336
|
| 27 |
+
}
|
| 28 |
+
}
|
devkit/processor_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"image_token": "<image>",
|
| 3 |
+
"num_additional_image_tokens": 1,
|
| 4 |
+
"patch_size": 14,
|
| 5 |
+
"processor_class": "LlavaProcessor",
|
| 6 |
+
"vision_feature_select_strategy": "default"
|
| 7 |
+
}
|
devkit/special_tokens_map.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"eos_token": {
|
| 10 |
+
"content": "</s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "<pad>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"unk_token": {
|
| 24 |
+
"content": "<unk>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
}
|
| 30 |
+
}
|
devkit/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
devkit/tokenizer_config.json
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"add_prefix_space": null,
|
| 5 |
+
"added_tokens_decoder": {
|
| 6 |
+
"0": {
|
| 7 |
+
"content": "<unk>",
|
| 8 |
+
"lstrip": false,
|
| 9 |
+
"normalized": false,
|
| 10 |
+
"rstrip": false,
|
| 11 |
+
"single_word": false,
|
| 12 |
+
"special": true
|
| 13 |
+
},
|
| 14 |
+
"1": {
|
| 15 |
+
"content": "<s>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false,
|
| 20 |
+
"special": true
|
| 21 |
+
},
|
| 22 |
+
"2": {
|
| 23 |
+
"content": "</s>",
|
| 24 |
+
"lstrip": false,
|
| 25 |
+
"normalized": false,
|
| 26 |
+
"rstrip": false,
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"special": true
|
| 29 |
+
},
|
| 30 |
+
"32000": {
|
| 31 |
+
"content": "<image>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false,
|
| 36 |
+
"special": true
|
| 37 |
+
},
|
| 38 |
+
"32001": {
|
| 39 |
+
"content": "<pad>",
|
| 40 |
+
"lstrip": false,
|
| 41 |
+
"normalized": false,
|
| 42 |
+
"rstrip": false,
|
| 43 |
+
"single_word": false,
|
| 44 |
+
"special": true
|
| 45 |
+
}
|
| 46 |
+
},
|
| 47 |
+
"bos_token": "<s>",
|
| 48 |
+
"clean_up_tokenization_spaces": false,
|
| 49 |
+
"eos_token": "</s>",
|
| 50 |
+
"extra_special_tokens": {
|
| 51 |
+
"image_token": "<image>"
|
| 52 |
+
},
|
| 53 |
+
"image_token": "<image>",
|
| 54 |
+
"legacy": false,
|
| 55 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 56 |
+
"pad_token": "<pad>",
|
| 57 |
+
"padding_side": "left",
|
| 58 |
+
"processor_class": "LlavaProcessor",
|
| 59 |
+
"sp_model_kwargs": {},
|
| 60 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 61 |
+
"trust_remote_code": false,
|
| 62 |
+
"unk_token": "<unk>",
|
| 63 |
+
"use_default_system_prompt": false
|
| 64 |
+
}
|
devkit/vlm_config.json
CHANGED
|
@@ -27,14 +27,7 @@
|
|
| 27 |
"gen": "2",
|
| 28 |
"size": "7b",
|
| 29 |
"token_cfg": {
|
| 30 |
-
"
|
| 31 |
-
"tokenizer_path": "/project/mlasw/users/amy.chen/debug/models--llava-hf--llava-1.5-7b-hf/blobs/9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347",
|
| 32 |
-
"vocab_size": 32064,
|
| 33 |
-
"special_tokens": {
|
| 34 |
-
"ignore_index": -100,
|
| 35 |
-
"image_token_index": 32000,
|
| 36 |
-
"pad_token_id": 32001
|
| 37 |
-
}
|
| 38 |
},
|
| 39 |
"rope_cfg": {
|
| 40 |
"rope_theta": 10000,
|
|
@@ -70,6 +63,7 @@
|
|
| 70 |
"num_hidden_layers": 32,
|
| 71 |
"max_position_embeddings": 2048,
|
| 72 |
"rms_norm_eps": 1e-05,
|
|
|
|
| 73 |
"layer_norms": [
|
| 74 |
"pre_attn",
|
| 75 |
"post_attn"
|
|
@@ -80,8 +74,8 @@
|
|
| 80 |
"lm_head_split_dim": 32064
|
| 81 |
},
|
| 82 |
"pipeline_cfg": {
|
| 83 |
-
"system_prompt":
|
| 84 |
-
"max_num_tokens":
|
| 85 |
"input_token_group_size": 100,
|
| 86 |
"input_token_group_offsets": [
|
| 87 |
0,
|
|
@@ -90,9 +84,27 @@
|
|
| 90 |
235,
|
| 91 |
335,
|
| 92 |
435,
|
| 93 |
-
535
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 94 |
],
|
| 95 |
-
"future_token_mask_size": 128
|
|
|
|
|
|
|
|
|
|
| 96 |
},
|
| 97 |
"language_model_name": "models--llava-hf--llava-1.5-7b-hf_language",
|
| 98 |
"vision_model_name": "models--llava-hf--llava-1.5-7b-hf_vision"
|
|
|
|
| 27 |
"gen": "2",
|
| 28 |
"size": "7b",
|
| 29 |
"token_cfg": {
|
| 30 |
+
"vocab_size": 32064
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
},
|
| 32 |
"rope_cfg": {
|
| 33 |
"rope_theta": 10000,
|
|
|
|
| 63 |
"num_hidden_layers": 32,
|
| 64 |
"max_position_embeddings": 2048,
|
| 65 |
"rms_norm_eps": 1e-05,
|
| 66 |
+
"rms_norm_unit_offset": false,
|
| 67 |
"layer_norms": [
|
| 68 |
"pre_attn",
|
| 69 |
"post_attn"
|
|
|
|
| 74 |
"lm_head_split_dim": 32064
|
| 75 |
},
|
| 76 |
"pipeline_cfg": {
|
| 77 |
+
"system_prompt": "A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.",
|
| 78 |
+
"max_num_tokens": 2048,
|
| 79 |
"input_token_group_size": 100,
|
| 80 |
"input_token_group_offsets": [
|
| 81 |
0,
|
|
|
|
| 84 |
235,
|
| 85 |
335,
|
| 86 |
435,
|
| 87 |
+
535,
|
| 88 |
+
635,
|
| 89 |
+
735,
|
| 90 |
+
835,
|
| 91 |
+
935,
|
| 92 |
+
1035,
|
| 93 |
+
1135,
|
| 94 |
+
1235,
|
| 95 |
+
1335,
|
| 96 |
+
1435,
|
| 97 |
+
1535,
|
| 98 |
+
1635,
|
| 99 |
+
1735,
|
| 100 |
+
1835,
|
| 101 |
+
1935,
|
| 102 |
+
2035
|
| 103 |
],
|
| 104 |
+
"future_token_mask_size": 128,
|
| 105 |
+
"return_logits": false,
|
| 106 |
+
"use_strided_kv_cache": false,
|
| 107 |
+
"enable_filter_sharing": true
|
| 108 |
},
|
| 109 |
"language_model_name": "models--llava-hf--llava-1.5-7b-hf_language",
|
| 110 |
"vision_model_name": "models--llava-hf--llava-1.5-7b-hf_vision"
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token0_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cb05f04908cf4620d317f8ae37c233d1649637b4b570847b03f72d824f600bec
|
| 3 |
+
size 5249384
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1035_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b43cc2290fd06448d5e3f7bc97be81da2b488edb92e8c1ab65277f81653cca2e
|
| 3 |
+
size 16441184
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1135_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be346dc5ec03e9d0e8fd61d1486694bf6ae4296ddfac4694f231c34b6032d075
|
| 3 |
+
size 26562424
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1235_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d36094e1bac50743b5d9f4b86a124da7fcf5346b841c9596eec550ab44ddb2e
|
| 3 |
+
size 25794480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1335_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40e66f5b3a3a64c32f384a50013bdcb4680082a32ef988d29b7d045fa30e7d20
|
| 3 |
+
size 26602576
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token135_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:674bd8ce85fd6901b851f2db567f8af35ad36592789fd392e58ad93eb26d1ded
|
| 3 |
+
size 5930936
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1435_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58e4f204c913457d031e0c8f00aa9d65c7f02e07529eab71b3aae6d49f237aeb
|
| 3 |
+
size 26046800
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1535_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be5c3b08a0f0c74acdc9cd9a379a428496920f8b91e54efca215ee0396e42abf
|
| 3 |
+
size 39036456
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1635_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06e201d2a209d8bb8f5b495646c10bf4efa47ab76de99ab4297b4df48b5afcaf
|
| 3 |
+
size 38202128
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1735_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bbc611431a5344a4ad29350c347cc12aefafb033bc7ff22667d748e0dfb58e93
|
| 3 |
+
size 27141312
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1835_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ef9a13a88fedfe41da8442ac6b9564b3c97b51ccc674498e1f83d8ee589c13a2
|
| 3 |
+
size 46105152
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token1935_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4342ecc45a9dd04096349f92db562804674f683bb1a0e3bc44f41cf463eb60a1
|
| 3 |
+
size 42110864
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token2035_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eed8e399c9d938b55f784d4f9b8076c31a79f2a78b3375971a4db28b617a405d
|
| 3 |
+
size 46698360
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token235_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ee238a7572d856899986ae1304ab105ae7e2ea9ea8fcf00a87b78f519d7f70c
|
| 3 |
+
size 6523544
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token335_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:099cec4a101f802df76ca520f6f15b6e88ef2f618721dc915d8867a363d8b6af
|
| 3 |
+
size 6831208
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token35_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:488c194be10d87ef397a6d49860b5eae149afc9f452640d335aa525c4cb3efe0
|
| 3 |
+
size 5726968
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token435_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d55faabe3956c16c723fff3330ae40c90685aad30ea25f7722ec29651cbb2470
|
| 3 |
+
size 10372416
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token535_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6e9707275e39f88867e75aaf4d2af5d8385a88a47864562d4bfee1d735ce4de9
|
| 3 |
+
size 9839976
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token635_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ec61d5df208bc9da15ca86b8127d4841d99c67db9673869823ec4097ec42e64
|
| 3 |
+
size 14145048
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token735_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:20887c5dfbb76be03f0121ace6436e482051beb72a6f4f4a56c8cc4309f798ff
|
| 3 |
+
size 13813688
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token835_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2e8da5055903211c589b2626c0f318c4f51a2874d4abca9b6aa6b362ed6bf6a2
|
| 3 |
+
size 19951320
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_cache_token935_stage1_mla.elf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:30e8b75b2870c5e997dfd659dcc9027302d73452ecc65cad3ba4994fceaee442
|
| 3 |
+
size 16231720
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer0_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c072579e27edd852f4d2424c6281ccea71c6e6cebda339007637c9a22f2a4e9d
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer10_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db0ed83b278fdfb42df3811f03c30da4daf6500febec5e001234ede493718c6c
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer11_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be876348c84c18c9a53fe9dde3f971cac04c7a55d0a2df27fa07ab9bc40bd26d
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer12_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2352c7f4989babd52f8e36e744912e690a873b81b6125351bfa6756879943e4
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer13_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:294039b4338d048f83f6fc1d177b2de95376f0ffb41408d0d062efa822afdc9e
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer14_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1885d3d13bac0ce3069df6c6caa11a44c7960bd0418917ce44f25362a572bf7a
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer15_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93363e79b47830f8db10e8252ea3fdd3c511c67d3a42abe14fb93c341df9787d
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer16_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb365a30d2745a7c91fddc91e6735a5f3227ceb5ac5045918b694156705d84a2
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer17_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cdc6c0f4908863393766dde05e9f922fd39b60f8628f9406508ae36f93eeba40
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer18_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ccde352c1c479d733b2d5f18005f86382d540a1397a49369a80c32333e7e0377
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer19_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:57e9a88cf35d053d47e884e79fe3ce5263a6bd8f6f152918bf163e490eeae8a6
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer1_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f02da00bbe9d6c6865d6883f8d1b396b59a17b1809ad2a087ddc9f6670092ebb
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer20_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:61fe1f5464934e983446c43f4ae19921b29213583113e83df1af0498cb5233ba
|
| 3 |
+
size 204496480
|
elf_files/models--llava-hf--llava-1.5-7b-hf_language_n100_post_layer21_stage1_mla.elf
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5c5804366a16bb199b27e14177313270223b4967c78599adf8f9679c8cd06f4
|
| 3 |
+
size 204496480
|