Hugging Face's logo

NeoChen1024
/
llama-joycaption-beta-one-hf-llava-FP8-Dynamic

Safetensors
llava
captioning
compressed-tensors
Model card Files Community
NeoChen1024 commited on
Commit
fdcedfe
·
verified ·
1 Parent(s): 8e9b506

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +166 -0
  2. model.safetensors.index.json +0 -0
  3. recipe.yaml +1 -1
config.json CHANGED
@@ -45,6 +45,172 @@
45
  "format": "float-quantized",
46
  "global_compression_ratio": null,
47
  "ignore": [
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
48
  "lm_head"
49
  ],
50
  "kv_cache_scheme": null,
 
45
  "format": "float-quantized",
46
  "global_compression_ratio": null,
47
  "ignore": [
48
+ "model.vision_tower.vision_model.encoder.layers.0.self_attn.k_proj",
49
+ "model.vision_tower.vision_model.encoder.layers.0.self_attn.v_proj",
50
+ "model.vision_tower.vision_model.encoder.layers.0.self_attn.q_proj",
51
+ "model.vision_tower.vision_model.encoder.layers.0.self_attn.out_proj",
52
+ "model.vision_tower.vision_model.encoder.layers.0.mlp.fc1",
53
+ "model.vision_tower.vision_model.encoder.layers.0.mlp.fc2",
54
+ "model.vision_tower.vision_model.encoder.layers.1.self_attn.k_proj",
55
+ "model.vision_tower.vision_model.encoder.layers.1.self_attn.v_proj",
56
+ "model.vision_tower.vision_model.encoder.layers.1.self_attn.q_proj",
57
+ "model.vision_tower.vision_model.encoder.layers.1.self_attn.out_proj",
58
+ "model.vision_tower.vision_model.encoder.layers.1.mlp.fc1",
59
+ "model.vision_tower.vision_model.encoder.layers.1.mlp.fc2",
60
+ "model.vision_tower.vision_model.encoder.layers.2.self_attn.k_proj",
61
+ "model.vision_tower.vision_model.encoder.layers.2.self_attn.v_proj",
62
+ "model.vision_tower.vision_model.encoder.layers.2.self_attn.q_proj",
63
+ "model.vision_tower.vision_model.encoder.layers.2.self_attn.out_proj",
64
+ "model.vision_tower.vision_model.encoder.layers.2.mlp.fc1",
65
+ "model.vision_tower.vision_model.encoder.layers.2.mlp.fc2",
66
+ "model.vision_tower.vision_model.encoder.layers.3.self_attn.k_proj",
67
+ "model.vision_tower.vision_model.encoder.layers.3.self_attn.v_proj",
68
+ "model.vision_tower.vision_model.encoder.layers.3.self_attn.q_proj",
69
+ "model.vision_tower.vision_model.encoder.layers.3.self_attn.out_proj",
70
+ "model.vision_tower.vision_model.encoder.layers.3.mlp.fc1",
71
+ "model.vision_tower.vision_model.encoder.layers.3.mlp.fc2",
72
+ "model.vision_tower.vision_model.encoder.layers.4.self_attn.k_proj",
73
+ "model.vision_tower.vision_model.encoder.layers.4.self_attn.v_proj",
74
+ "model.vision_tower.vision_model.encoder.layers.4.self_attn.q_proj",
75
+ "model.vision_tower.vision_model.encoder.layers.4.self_attn.out_proj",
76
+ "model.vision_tower.vision_model.encoder.layers.4.mlp.fc1",
77
+ "model.vision_tower.vision_model.encoder.layers.4.mlp.fc2",
78
+ "model.vision_tower.vision_model.encoder.layers.5.self_attn.k_proj",
79
+ "model.vision_tower.vision_model.encoder.layers.5.self_attn.v_proj",
80
+ "model.vision_tower.vision_model.encoder.layers.5.self_attn.q_proj",
81
+ "model.vision_tower.vision_model.encoder.layers.5.self_attn.out_proj",
82
+ "model.vision_tower.vision_model.encoder.layers.5.mlp.fc1",
83
+ "model.vision_tower.vision_model.encoder.layers.5.mlp.fc2",
84
+ "model.vision_tower.vision_model.encoder.layers.6.self_attn.k_proj",
85
+ "model.vision_tower.vision_model.encoder.layers.6.self_attn.v_proj",
86
+ "model.vision_tower.vision_model.encoder.layers.6.self_attn.q_proj",
87
+ "model.vision_tower.vision_model.encoder.layers.6.self_attn.out_proj",
88
+ "model.vision_tower.vision_model.encoder.layers.6.mlp.fc1",
89
+ "model.vision_tower.vision_model.encoder.layers.6.mlp.fc2",
90
+ "model.vision_tower.vision_model.encoder.layers.7.self_attn.k_proj",
91
+ "model.vision_tower.vision_model.encoder.layers.7.self_attn.v_proj",
92
+ "model.vision_tower.vision_model.encoder.layers.7.self_attn.q_proj",
93
+ "model.vision_tower.vision_model.encoder.layers.7.self_attn.out_proj",
94
+ "model.vision_tower.vision_model.encoder.layers.7.mlp.fc1",
95
+ "model.vision_tower.vision_model.encoder.layers.7.mlp.fc2",
96
+ "model.vision_tower.vision_model.encoder.layers.8.self_attn.k_proj",
97
+ "model.vision_tower.vision_model.encoder.layers.8.self_attn.v_proj",
98
+ "model.vision_tower.vision_model.encoder.layers.8.self_attn.q_proj",
99
+ "model.vision_tower.vision_model.encoder.layers.8.self_attn.out_proj",
100
+ "model.vision_tower.vision_model.encoder.layers.8.mlp.fc1",
101
+ "model.vision_tower.vision_model.encoder.layers.8.mlp.fc2",
102
+ "model.vision_tower.vision_model.encoder.layers.9.self_attn.k_proj",
103
+ "model.vision_tower.vision_model.encoder.layers.9.self_attn.v_proj",
104
+ "model.vision_tower.vision_model.encoder.layers.9.self_attn.q_proj",
105
+ "model.vision_tower.vision_model.encoder.layers.9.self_attn.out_proj",
106
+ "model.vision_tower.vision_model.encoder.layers.9.mlp.fc1",
107
+ "model.vision_tower.vision_model.encoder.layers.9.mlp.fc2",
108
+ "model.vision_tower.vision_model.encoder.layers.10.self_attn.k_proj",
109
+ "model.vision_tower.vision_model.encoder.layers.10.self_attn.v_proj",
110
+ "model.vision_tower.vision_model.encoder.layers.10.self_attn.q_proj",
111
+ "model.vision_tower.vision_model.encoder.layers.10.self_attn.out_proj",
112
+ "model.vision_tower.vision_model.encoder.layers.10.mlp.fc1",
113
+ "model.vision_tower.vision_model.encoder.layers.10.mlp.fc2",
114
+ "model.vision_tower.vision_model.encoder.layers.11.self_attn.k_proj",
115
+ "model.vision_tower.vision_model.encoder.layers.11.self_attn.v_proj",
116
+ "model.vision_tower.vision_model.encoder.layers.11.self_attn.q_proj",
117
+ "model.vision_tower.vision_model.encoder.layers.11.self_attn.out_proj",
118
+ "model.vision_tower.vision_model.encoder.layers.11.mlp.fc1",
119
+ "model.vision_tower.vision_model.encoder.layers.11.mlp.fc2",
120
+ "model.vision_tower.vision_model.encoder.layers.12.self_attn.k_proj",
121
+ "model.vision_tower.vision_model.encoder.layers.12.self_attn.v_proj",
122
+ "model.vision_tower.vision_model.encoder.layers.12.self_attn.q_proj",
123
+ "model.vision_tower.vision_model.encoder.layers.12.self_attn.out_proj",
124
+ "model.vision_tower.vision_model.encoder.layers.12.mlp.fc1",
125
+ "model.vision_tower.vision_model.encoder.layers.12.mlp.fc2",
126
+ "model.vision_tower.vision_model.encoder.layers.13.self_attn.k_proj",
127
+ "model.vision_tower.vision_model.encoder.layers.13.self_attn.v_proj",
128
+ "model.vision_tower.vision_model.encoder.layers.13.self_attn.q_proj",
129
+ "model.vision_tower.vision_model.encoder.layers.13.self_attn.out_proj",
130
+ "model.vision_tower.vision_model.encoder.layers.13.mlp.fc1",
131
+ "model.vision_tower.vision_model.encoder.layers.13.mlp.fc2",
132
+ "model.vision_tower.vision_model.encoder.layers.14.self_attn.k_proj",
133
+ "model.vision_tower.vision_model.encoder.layers.14.self_attn.v_proj",
134
+ "model.vision_tower.vision_model.encoder.layers.14.self_attn.q_proj",
135
+ "model.vision_tower.vision_model.encoder.layers.14.self_attn.out_proj",
136
+ "model.vision_tower.vision_model.encoder.layers.14.mlp.fc1",
137
+ "model.vision_tower.vision_model.encoder.layers.14.mlp.fc2",
138
+ "model.vision_tower.vision_model.encoder.layers.15.self_attn.k_proj",
139
+ "model.vision_tower.vision_model.encoder.layers.15.self_attn.v_proj",
140
+ "model.vision_tower.vision_model.encoder.layers.15.self_attn.q_proj",
141
+ "model.vision_tower.vision_model.encoder.layers.15.self_attn.out_proj",
142
+ "model.vision_tower.vision_model.encoder.layers.15.mlp.fc1",
143
+ "model.vision_tower.vision_model.encoder.layers.15.mlp.fc2",
144
+ "model.vision_tower.vision_model.encoder.layers.16.self_attn.k_proj",
145
+ "model.vision_tower.vision_model.encoder.layers.16.self_attn.v_proj",
146
+ "model.vision_tower.vision_model.encoder.layers.16.self_attn.q_proj",
147
+ "model.vision_tower.vision_model.encoder.layers.16.self_attn.out_proj",
148
+ "model.vision_tower.vision_model.encoder.layers.16.mlp.fc1",
149
+ "model.vision_tower.vision_model.encoder.layers.16.mlp.fc2",
150
+ "model.vision_tower.vision_model.encoder.layers.17.self_attn.k_proj",
151
+ "model.vision_tower.vision_model.encoder.layers.17.self_attn.v_proj",
152
+ "model.vision_tower.vision_model.encoder.layers.17.self_attn.q_proj",
153
+ "model.vision_tower.vision_model.encoder.layers.17.self_attn.out_proj",
154
+ "model.vision_tower.vision_model.encoder.layers.17.mlp.fc1",
155
+ "model.vision_tower.vision_model.encoder.layers.17.mlp.fc2",
156
+ "model.vision_tower.vision_model.encoder.layers.18.self_attn.k_proj",
157
+ "model.vision_tower.vision_model.encoder.layers.18.self_attn.v_proj",
158
+ "model.vision_tower.vision_model.encoder.layers.18.self_attn.q_proj",
159
+ "model.vision_tower.vision_model.encoder.layers.18.self_attn.out_proj",
160
+ "model.vision_tower.vision_model.encoder.layers.18.mlp.fc1",
161
+ "model.vision_tower.vision_model.encoder.layers.18.mlp.fc2",
162
+ "model.vision_tower.vision_model.encoder.layers.19.self_attn.k_proj",
163
+ "model.vision_tower.vision_model.encoder.layers.19.self_attn.v_proj",
164
+ "model.vision_tower.vision_model.encoder.layers.19.self_attn.q_proj",
165
+ "model.vision_tower.vision_model.encoder.layers.19.self_attn.out_proj",
166
+ "model.vision_tower.vision_model.encoder.layers.19.mlp.fc1",
167
+ "model.vision_tower.vision_model.encoder.layers.19.mlp.fc2",
168
+ "model.vision_tower.vision_model.encoder.layers.20.self_attn.k_proj",
169
+ "model.vision_tower.vision_model.encoder.layers.20.self_attn.v_proj",
170
+ "model.vision_tower.vision_model.encoder.layers.20.self_attn.q_proj",
171
+ "model.vision_tower.vision_model.encoder.layers.20.self_attn.out_proj",
172
+ "model.vision_tower.vision_model.encoder.layers.20.mlp.fc1",
173
+ "model.vision_tower.vision_model.encoder.layers.20.mlp.fc2",
174
+ "model.vision_tower.vision_model.encoder.layers.21.self_attn.k_proj",
175
+ "model.vision_tower.vision_model.encoder.layers.21.self_attn.v_proj",
176
+ "model.vision_tower.vision_model.encoder.layers.21.self_attn.q_proj",
177
+ "model.vision_tower.vision_model.encoder.layers.21.self_attn.out_proj",
178
+ "model.vision_tower.vision_model.encoder.layers.21.mlp.fc1",
179
+ "model.vision_tower.vision_model.encoder.layers.21.mlp.fc2",
180
+ "model.vision_tower.vision_model.encoder.layers.22.self_attn.k_proj",
181
+ "model.vision_tower.vision_model.encoder.layers.22.self_attn.v_proj",
182
+ "model.vision_tower.vision_model.encoder.layers.22.self_attn.q_proj",
183
+ "model.vision_tower.vision_model.encoder.layers.22.self_attn.out_proj",
184
+ "model.vision_tower.vision_model.encoder.layers.22.mlp.fc1",
185
+ "model.vision_tower.vision_model.encoder.layers.22.mlp.fc2",
186
+ "model.vision_tower.vision_model.encoder.layers.23.self_attn.k_proj",
187
+ "model.vision_tower.vision_model.encoder.layers.23.self_attn.v_proj",
188
+ "model.vision_tower.vision_model.encoder.layers.23.self_attn.q_proj",
189
+ "model.vision_tower.vision_model.encoder.layers.23.self_attn.out_proj",
190
+ "model.vision_tower.vision_model.encoder.layers.23.mlp.fc1",
191
+ "model.vision_tower.vision_model.encoder.layers.23.mlp.fc2",
192
+ "model.vision_tower.vision_model.encoder.layers.24.self_attn.k_proj",
193
+ "model.vision_tower.vision_model.encoder.layers.24.self_attn.v_proj",
194
+ "model.vision_tower.vision_model.encoder.layers.24.self_attn.q_proj",
195
+ "model.vision_tower.vision_model.encoder.layers.24.self_attn.out_proj",
196
+ "model.vision_tower.vision_model.encoder.layers.24.mlp.fc1",
197
+ "model.vision_tower.vision_model.encoder.layers.24.mlp.fc2",
198
+ "model.vision_tower.vision_model.encoder.layers.25.self_attn.k_proj",
199
+ "model.vision_tower.vision_model.encoder.layers.25.self_attn.v_proj",
200
+ "model.vision_tower.vision_model.encoder.layers.25.self_attn.q_proj",
201
+ "model.vision_tower.vision_model.encoder.layers.25.self_attn.out_proj",
202
+ "model.vision_tower.vision_model.encoder.layers.25.mlp.fc1",
203
+ "model.vision_tower.vision_model.encoder.layers.25.mlp.fc2",
204
+ "model.vision_tower.vision_model.encoder.layers.26.self_attn.k_proj",
205
+ "model.vision_tower.vision_model.encoder.layers.26.self_attn.v_proj",
206
+ "model.vision_tower.vision_model.encoder.layers.26.self_attn.q_proj",
207
+ "model.vision_tower.vision_model.encoder.layers.26.self_attn.out_proj",
208
+ "model.vision_tower.vision_model.encoder.layers.26.mlp.fc1",
209
+ "model.vision_tower.vision_model.encoder.layers.26.mlp.fc2",
210
+ "model.vision_tower.vision_model.head.mlp.fc1",
211
+ "model.vision_tower.vision_model.head.mlp.fc2",
212
+ "model.multi_modal_projector.linear_1",
213
+ "model.multi_modal_projector.linear_2",
214
  "lm_head"
215
  ],
216
  "kv_cache_scheme": null,
model.safetensors.index.json CHANGED
The diff for this file is too large to render. See raw diff
 
recipe.yaml CHANGED
@@ -2,5 +2,5 @@ default_stage:
2
  default_modifiers:
3
  QuantizationModifier:
4
  targets: [Linear]
5
- ignore: ['re:.*lm_head', 're:multi_modal_projector.*', 're:vision_tower.*']
6
  scheme: FP8_DYNAMIC
 
2
  default_modifiers:
3
  QuantizationModifier:
4
  targets: [Linear]
5
+ ignore: ['re:.*lm_head', 're:.*multi_modal_projector.*', 're:.*vision_tower.*']
6
  scheme: FP8_DYNAMIC