Shubhra Pandit
commited on
Commit
•
2deae6b
1
Parent(s):
1d782bd
Upload model files
Browse files- .gitattributes +1 -0
- README.md +0 -3
- chat_template.json +3 -0
- config.json +700 -0
- generation_config.json +6 -0
- model-00001-of-00002.safetensors +3 -0
- model-00002-of-00002.safetensors +3 -0
- model.safetensors.index.json +0 -0
- preprocessor_config.json +27 -0
- processor_config.json +7 -0
- recipe.yaml +11 -0
- special_tokens_map.json +24 -0
- tokenizer.json +3 -0
- tokenizer_config.json +0 -0
.gitattributes
CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
---
|
2 |
-
license: apache-2.0
|
3 |
-
---
|
|
|
|
|
|
|
|
chat_template.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"chat_template": "{%- if messages[0][\"role\"] == \"system\" %}\n {%- set system_message = messages[0][\"content\"] %}\n {%- set loop_messages = messages[1:] %}\n{%- else %}\n {%- set loop_messages = messages %}\n{%- endif %}\n\n{{- bos_token }}\n{%- for message in loop_messages %}\n {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}\n {{- raise_exception('After the optional system message, conversation roles must alternate user/assistant/user/assistant/...') }}\n {%- endif %}\n {%- if message[\"role\"] == \"user\" %}\n {%- if loop.last and system_message is defined %}\n {{- \"[INST]\" + system_message + \"\n\n\" }}\n {%- else %}\n {{- \"[INST]\" }}\n {%- endif %}\n {%- if message[\"content\"] is not string %}\n {%- for chunk in message[\"content\"] %}\n {%- if chunk[\"type\"] == \"text\" %}\n {{- chunk[\"content\"] }}\n {%- elif chunk[\"type\"] == \"image\" %}\n {{- \"[IMG]\" }}\n {%- else %}\n {{- raise_exception(\"Unrecognized content type!\") }}\n {%- endif %}\n {%- endfor %}\n {%- else %}\n {{- message[\"content\"] }}\n {%- endif %}\n {{- \"[/INST]\" }}\n {%- elif message[\"role\"] == \"assistant\" %}\n {{- message[\"content\"] + eos_token}}\n {%- else %}\n {{- raise_exception(\"Only user and assistant roles are supported, with the exception of an initial optional system message!\") }}\n {%- endif %}\n{%- endfor %}"
|
3 |
+
}
|
config.json
ADDED
@@ -0,0 +1,700 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "mgoin/pixtral-12b",
|
3 |
+
"architectures": [
|
4 |
+
"LlavaForConditionalGeneration"
|
5 |
+
],
|
6 |
+
"ignore_index": -100,
|
7 |
+
"image_seq_length": 1,
|
8 |
+
"image_token_index": 10,
|
9 |
+
"model_type": "llava",
|
10 |
+
"projector_hidden_act": "gelu",
|
11 |
+
"quantization_config": {
|
12 |
+
"config_groups": {
|
13 |
+
"group_0": {
|
14 |
+
"input_activations": null,
|
15 |
+
"output_activations": null,
|
16 |
+
"targets": [
|
17 |
+
"Linear"
|
18 |
+
],
|
19 |
+
"weights": {
|
20 |
+
"actorder": "weight",
|
21 |
+
"block_structure": null,
|
22 |
+
"dynamic": false,
|
23 |
+
"group_size": 128,
|
24 |
+
"num_bits": 4,
|
25 |
+
"observer": "minmax",
|
26 |
+
"observer_kwargs": {},
|
27 |
+
"strategy": "group",
|
28 |
+
"symmetric": true,
|
29 |
+
"type": "int"
|
30 |
+
}
|
31 |
+
}
|
32 |
+
},
|
33 |
+
"format": "pack-quantized",
|
34 |
+
"global_compression_ratio": 1.4377147085920574,
|
35 |
+
"ignore": [
|
36 |
+
"vision_tower.transformer.layers.0.feed_forward.gate_proj",
|
37 |
+
"vision_tower.transformer.layers.0.feed_forward.up_proj",
|
38 |
+
"vision_tower.transformer.layers.0.feed_forward.down_proj",
|
39 |
+
"vision_tower.transformer.layers.0.attention.k_proj",
|
40 |
+
"vision_tower.transformer.layers.0.attention.v_proj",
|
41 |
+
"vision_tower.transformer.layers.0.attention.q_proj",
|
42 |
+
"vision_tower.transformer.layers.0.attention.o_proj",
|
43 |
+
"vision_tower.transformer.layers.1.feed_forward.gate_proj",
|
44 |
+
"vision_tower.transformer.layers.1.feed_forward.up_proj",
|
45 |
+
"vision_tower.transformer.layers.1.feed_forward.down_proj",
|
46 |
+
"vision_tower.transformer.layers.1.attention.k_proj",
|
47 |
+
"vision_tower.transformer.layers.1.attention.v_proj",
|
48 |
+
"vision_tower.transformer.layers.1.attention.q_proj",
|
49 |
+
"vision_tower.transformer.layers.1.attention.o_proj",
|
50 |
+
"vision_tower.transformer.layers.2.feed_forward.gate_proj",
|
51 |
+
"vision_tower.transformer.layers.2.feed_forward.up_proj",
|
52 |
+
"vision_tower.transformer.layers.2.feed_forward.down_proj",
|
53 |
+
"vision_tower.transformer.layers.2.attention.k_proj",
|
54 |
+
"vision_tower.transformer.layers.2.attention.v_proj",
|
55 |
+
"vision_tower.transformer.layers.2.attention.q_proj",
|
56 |
+
"vision_tower.transformer.layers.2.attention.o_proj",
|
57 |
+
"vision_tower.transformer.layers.3.feed_forward.gate_proj",
|
58 |
+
"vision_tower.transformer.layers.3.feed_forward.up_proj",
|
59 |
+
"vision_tower.transformer.layers.3.feed_forward.down_proj",
|
60 |
+
"vision_tower.transformer.layers.3.attention.k_proj",
|
61 |
+
"vision_tower.transformer.layers.3.attention.v_proj",
|
62 |
+
"vision_tower.transformer.layers.3.attention.q_proj",
|
63 |
+
"vision_tower.transformer.layers.3.attention.o_proj",
|
64 |
+
"vision_tower.transformer.layers.4.feed_forward.gate_proj",
|
65 |
+
"vision_tower.transformer.layers.4.feed_forward.up_proj",
|
66 |
+
"vision_tower.transformer.layers.4.feed_forward.down_proj",
|
67 |
+
"vision_tower.transformer.layers.4.attention.k_proj",
|
68 |
+
"vision_tower.transformer.layers.4.attention.v_proj",
|
69 |
+
"vision_tower.transformer.layers.4.attention.q_proj",
|
70 |
+
"vision_tower.transformer.layers.4.attention.o_proj",
|
71 |
+
"vision_tower.transformer.layers.5.feed_forward.gate_proj",
|
72 |
+
"vision_tower.transformer.layers.5.feed_forward.up_proj",
|
73 |
+
"vision_tower.transformer.layers.5.feed_forward.down_proj",
|
74 |
+
"vision_tower.transformer.layers.5.attention.k_proj",
|
75 |
+
"vision_tower.transformer.layers.5.attention.v_proj",
|
76 |
+
"vision_tower.transformer.layers.5.attention.q_proj",
|
77 |
+
"vision_tower.transformer.layers.5.attention.o_proj",
|
78 |
+
"vision_tower.transformer.layers.6.feed_forward.gate_proj",
|
79 |
+
"vision_tower.transformer.layers.6.feed_forward.up_proj",
|
80 |
+
"vision_tower.transformer.layers.6.feed_forward.down_proj",
|
81 |
+
"vision_tower.transformer.layers.6.attention.k_proj",
|
82 |
+
"vision_tower.transformer.layers.6.attention.v_proj",
|
83 |
+
"vision_tower.transformer.layers.6.attention.q_proj",
|
84 |
+
"vision_tower.transformer.layers.6.attention.o_proj",
|
85 |
+
"vision_tower.transformer.layers.7.feed_forward.gate_proj",
|
86 |
+
"vision_tower.transformer.layers.7.feed_forward.up_proj",
|
87 |
+
"vision_tower.transformer.layers.7.feed_forward.down_proj",
|
88 |
+
"vision_tower.transformer.layers.7.attention.k_proj",
|
89 |
+
"vision_tower.transformer.layers.7.attention.v_proj",
|
90 |
+
"vision_tower.transformer.layers.7.attention.q_proj",
|
91 |
+
"vision_tower.transformer.layers.7.attention.o_proj",
|
92 |
+
"vision_tower.transformer.layers.8.feed_forward.gate_proj",
|
93 |
+
"vision_tower.transformer.layers.8.feed_forward.up_proj",
|
94 |
+
"vision_tower.transformer.layers.8.feed_forward.down_proj",
|
95 |
+
"vision_tower.transformer.layers.8.attention.k_proj",
|
96 |
+
"vision_tower.transformer.layers.8.attention.v_proj",
|
97 |
+
"vision_tower.transformer.layers.8.attention.q_proj",
|
98 |
+
"vision_tower.transformer.layers.8.attention.o_proj",
|
99 |
+
"vision_tower.transformer.layers.9.feed_forward.gate_proj",
|
100 |
+
"vision_tower.transformer.layers.9.feed_forward.up_proj",
|
101 |
+
"vision_tower.transformer.layers.9.feed_forward.down_proj",
|
102 |
+
"vision_tower.transformer.layers.9.attention.k_proj",
|
103 |
+
"vision_tower.transformer.layers.9.attention.v_proj",
|
104 |
+
"vision_tower.transformer.layers.9.attention.q_proj",
|
105 |
+
"vision_tower.transformer.layers.9.attention.o_proj",
|
106 |
+
"vision_tower.transformer.layers.10.feed_forward.gate_proj",
|
107 |
+
"vision_tower.transformer.layers.10.feed_forward.up_proj",
|
108 |
+
"vision_tower.transformer.layers.10.feed_forward.down_proj",
|
109 |
+
"vision_tower.transformer.layers.10.attention.k_proj",
|
110 |
+
"vision_tower.transformer.layers.10.attention.v_proj",
|
111 |
+
"vision_tower.transformer.layers.10.attention.q_proj",
|
112 |
+
"vision_tower.transformer.layers.10.attention.o_proj",
|
113 |
+
"vision_tower.transformer.layers.11.feed_forward.gate_proj",
|
114 |
+
"vision_tower.transformer.layers.11.feed_forward.up_proj",
|
115 |
+
"vision_tower.transformer.layers.11.feed_forward.down_proj",
|
116 |
+
"vision_tower.transformer.layers.11.attention.k_proj",
|
117 |
+
"vision_tower.transformer.layers.11.attention.v_proj",
|
118 |
+
"vision_tower.transformer.layers.11.attention.q_proj",
|
119 |
+
"vision_tower.transformer.layers.11.attention.o_proj",
|
120 |
+
"vision_tower.transformer.layers.12.feed_forward.gate_proj",
|
121 |
+
"vision_tower.transformer.layers.12.feed_forward.up_proj",
|
122 |
+
"vision_tower.transformer.layers.12.feed_forward.down_proj",
|
123 |
+
"vision_tower.transformer.layers.12.attention.k_proj",
|
124 |
+
"vision_tower.transformer.layers.12.attention.v_proj",
|
125 |
+
"vision_tower.transformer.layers.12.attention.q_proj",
|
126 |
+
"vision_tower.transformer.layers.12.attention.o_proj",
|
127 |
+
"vision_tower.transformer.layers.13.feed_forward.gate_proj",
|
128 |
+
"vision_tower.transformer.layers.13.feed_forward.up_proj",
|
129 |
+
"vision_tower.transformer.layers.13.feed_forward.down_proj",
|
130 |
+
"vision_tower.transformer.layers.13.attention.k_proj",
|
131 |
+
"vision_tower.transformer.layers.13.attention.v_proj",
|
132 |
+
"vision_tower.transformer.layers.13.attention.q_proj",
|
133 |
+
"vision_tower.transformer.layers.13.attention.o_proj",
|
134 |
+
"vision_tower.transformer.layers.14.feed_forward.gate_proj",
|
135 |
+
"vision_tower.transformer.layers.14.feed_forward.up_proj",
|
136 |
+
"vision_tower.transformer.layers.14.feed_forward.down_proj",
|
137 |
+
"vision_tower.transformer.layers.14.attention.k_proj",
|
138 |
+
"vision_tower.transformer.layers.14.attention.v_proj",
|
139 |
+
"vision_tower.transformer.layers.14.attention.q_proj",
|
140 |
+
"vision_tower.transformer.layers.14.attention.o_proj",
|
141 |
+
"vision_tower.transformer.layers.15.feed_forward.gate_proj",
|
142 |
+
"vision_tower.transformer.layers.15.feed_forward.up_proj",
|
143 |
+
"vision_tower.transformer.layers.15.feed_forward.down_proj",
|
144 |
+
"vision_tower.transformer.layers.15.attention.k_proj",
|
145 |
+
"vision_tower.transformer.layers.15.attention.v_proj",
|
146 |
+
"vision_tower.transformer.layers.15.attention.q_proj",
|
147 |
+
"vision_tower.transformer.layers.15.attention.o_proj",
|
148 |
+
"vision_tower.transformer.layers.16.feed_forward.gate_proj",
|
149 |
+
"vision_tower.transformer.layers.16.feed_forward.up_proj",
|
150 |
+
"vision_tower.transformer.layers.16.feed_forward.down_proj",
|
151 |
+
"vision_tower.transformer.layers.16.attention.k_proj",
|
152 |
+
"vision_tower.transformer.layers.16.attention.v_proj",
|
153 |
+
"vision_tower.transformer.layers.16.attention.q_proj",
|
154 |
+
"vision_tower.transformer.layers.16.attention.o_proj",
|
155 |
+
"vision_tower.transformer.layers.17.feed_forward.gate_proj",
|
156 |
+
"vision_tower.transformer.layers.17.feed_forward.up_proj",
|
157 |
+
"vision_tower.transformer.layers.17.feed_forward.down_proj",
|
158 |
+
"vision_tower.transformer.layers.17.attention.k_proj",
|
159 |
+
"vision_tower.transformer.layers.17.attention.v_proj",
|
160 |
+
"vision_tower.transformer.layers.17.attention.q_proj",
|
161 |
+
"vision_tower.transformer.layers.17.attention.o_proj",
|
162 |
+
"vision_tower.transformer.layers.18.feed_forward.gate_proj",
|
163 |
+
"vision_tower.transformer.layers.18.feed_forward.up_proj",
|
164 |
+
"vision_tower.transformer.layers.18.feed_forward.down_proj",
|
165 |
+
"vision_tower.transformer.layers.18.attention.k_proj",
|
166 |
+
"vision_tower.transformer.layers.18.attention.v_proj",
|
167 |
+
"vision_tower.transformer.layers.18.attention.q_proj",
|
168 |
+
"vision_tower.transformer.layers.18.attention.o_proj",
|
169 |
+
"vision_tower.transformer.layers.19.feed_forward.gate_proj",
|
170 |
+
"vision_tower.transformer.layers.19.feed_forward.up_proj",
|
171 |
+
"vision_tower.transformer.layers.19.feed_forward.down_proj",
|
172 |
+
"vision_tower.transformer.layers.19.attention.k_proj",
|
173 |
+
"vision_tower.transformer.layers.19.attention.v_proj",
|
174 |
+
"vision_tower.transformer.layers.19.attention.q_proj",
|
175 |
+
"vision_tower.transformer.layers.19.attention.o_proj",
|
176 |
+
"vision_tower.transformer.layers.20.feed_forward.gate_proj",
|
177 |
+
"vision_tower.transformer.layers.20.feed_forward.up_proj",
|
178 |
+
"vision_tower.transformer.layers.20.feed_forward.down_proj",
|
179 |
+
"vision_tower.transformer.layers.20.attention.k_proj",
|
180 |
+
"vision_tower.transformer.layers.20.attention.v_proj",
|
181 |
+
"vision_tower.transformer.layers.20.attention.q_proj",
|
182 |
+
"vision_tower.transformer.layers.20.attention.o_proj",
|
183 |
+
"vision_tower.transformer.layers.21.feed_forward.gate_proj",
|
184 |
+
"vision_tower.transformer.layers.21.feed_forward.up_proj",
|
185 |
+
"vision_tower.transformer.layers.21.feed_forward.down_proj",
|
186 |
+
"vision_tower.transformer.layers.21.attention.k_proj",
|
187 |
+
"vision_tower.transformer.layers.21.attention.v_proj",
|
188 |
+
"vision_tower.transformer.layers.21.attention.q_proj",
|
189 |
+
"vision_tower.transformer.layers.21.attention.o_proj",
|
190 |
+
"vision_tower.transformer.layers.22.feed_forward.gate_proj",
|
191 |
+
"vision_tower.transformer.layers.22.feed_forward.up_proj",
|
192 |
+
"vision_tower.transformer.layers.22.feed_forward.down_proj",
|
193 |
+
"vision_tower.transformer.layers.22.attention.k_proj",
|
194 |
+
"vision_tower.transformer.layers.22.attention.v_proj",
|
195 |
+
"vision_tower.transformer.layers.22.attention.q_proj",
|
196 |
+
"vision_tower.transformer.layers.22.attention.o_proj",
|
197 |
+
"vision_tower.transformer.layers.23.feed_forward.gate_proj",
|
198 |
+
"vision_tower.transformer.layers.23.feed_forward.up_proj",
|
199 |
+
"vision_tower.transformer.layers.23.feed_forward.down_proj",
|
200 |
+
"vision_tower.transformer.layers.23.attention.k_proj",
|
201 |
+
"vision_tower.transformer.layers.23.attention.v_proj",
|
202 |
+
"vision_tower.transformer.layers.23.attention.q_proj",
|
203 |
+
"vision_tower.transformer.layers.23.attention.o_proj",
|
204 |
+
"multi_modal_projector.linear_1",
|
205 |
+
"multi_modal_projector.linear_2",
|
206 |
+
"language_model.lm_head"
|
207 |
+
],
|
208 |
+
"kv_cache_scheme": null,
|
209 |
+
"quant_method": "compressed-tensors",
|
210 |
+
"quantization_status": "compressed",
|
211 |
+
"sparsity_config": {
|
212 |
+
"format": "dense",
|
213 |
+
"global_sparsity": 0.13939651604852193,
|
214 |
+
"ignore": [
|
215 |
+
"vision_tower.transformer.layers.0.feed_forward.gate_proj",
|
216 |
+
"vision_tower.transformer.layers.0.feed_forward.up_proj",
|
217 |
+
"vision_tower.transformer.layers.0.feed_forward.down_proj",
|
218 |
+
"vision_tower.transformer.layers.0.attention.k_proj",
|
219 |
+
"vision_tower.transformer.layers.0.attention.v_proj",
|
220 |
+
"vision_tower.transformer.layers.0.attention.q_proj",
|
221 |
+
"vision_tower.transformer.layers.0.attention.o_proj",
|
222 |
+
"vision_tower.transformer.layers.1.feed_forward.gate_proj",
|
223 |
+
"vision_tower.transformer.layers.1.feed_forward.up_proj",
|
224 |
+
"vision_tower.transformer.layers.1.feed_forward.down_proj",
|
225 |
+
"vision_tower.transformer.layers.1.attention.k_proj",
|
226 |
+
"vision_tower.transformer.layers.1.attention.v_proj",
|
227 |
+
"vision_tower.transformer.layers.1.attention.q_proj",
|
228 |
+
"vision_tower.transformer.layers.1.attention.o_proj",
|
229 |
+
"vision_tower.transformer.layers.2.feed_forward.gate_proj",
|
230 |
+
"vision_tower.transformer.layers.2.feed_forward.up_proj",
|
231 |
+
"vision_tower.transformer.layers.2.feed_forward.down_proj",
|
232 |
+
"vision_tower.transformer.layers.2.attention.k_proj",
|
233 |
+
"vision_tower.transformer.layers.2.attention.v_proj",
|
234 |
+
"vision_tower.transformer.layers.2.attention.q_proj",
|
235 |
+
"vision_tower.transformer.layers.2.attention.o_proj",
|
236 |
+
"vision_tower.transformer.layers.3.feed_forward.gate_proj",
|
237 |
+
"vision_tower.transformer.layers.3.feed_forward.up_proj",
|
238 |
+
"vision_tower.transformer.layers.3.feed_forward.down_proj",
|
239 |
+
"vision_tower.transformer.layers.3.attention.k_proj",
|
240 |
+
"vision_tower.transformer.layers.3.attention.v_proj",
|
241 |
+
"vision_tower.transformer.layers.3.attention.q_proj",
|
242 |
+
"vision_tower.transformer.layers.3.attention.o_proj",
|
243 |
+
"vision_tower.transformer.layers.4.feed_forward.gate_proj",
|
244 |
+
"vision_tower.transformer.layers.4.feed_forward.up_proj",
|
245 |
+
"vision_tower.transformer.layers.4.feed_forward.down_proj",
|
246 |
+
"vision_tower.transformer.layers.4.attention.k_proj",
|
247 |
+
"vision_tower.transformer.layers.4.attention.v_proj",
|
248 |
+
"vision_tower.transformer.layers.4.attention.q_proj",
|
249 |
+
"vision_tower.transformer.layers.4.attention.o_proj",
|
250 |
+
"vision_tower.transformer.layers.5.feed_forward.gate_proj",
|
251 |
+
"vision_tower.transformer.layers.5.feed_forward.up_proj",
|
252 |
+
"vision_tower.transformer.layers.5.feed_forward.down_proj",
|
253 |
+
"vision_tower.transformer.layers.5.attention.k_proj",
|
254 |
+
"vision_tower.transformer.layers.5.attention.v_proj",
|
255 |
+
"vision_tower.transformer.layers.5.attention.q_proj",
|
256 |
+
"vision_tower.transformer.layers.5.attention.o_proj",
|
257 |
+
"vision_tower.transformer.layers.6.feed_forward.gate_proj",
|
258 |
+
"vision_tower.transformer.layers.6.feed_forward.up_proj",
|
259 |
+
"vision_tower.transformer.layers.6.feed_forward.down_proj",
|
260 |
+
"vision_tower.transformer.layers.6.attention.k_proj",
|
261 |
+
"vision_tower.transformer.layers.6.attention.v_proj",
|
262 |
+
"vision_tower.transformer.layers.6.attention.q_proj",
|
263 |
+
"vision_tower.transformer.layers.6.attention.o_proj",
|
264 |
+
"vision_tower.transformer.layers.7.feed_forward.gate_proj",
|
265 |
+
"vision_tower.transformer.layers.7.feed_forward.up_proj",
|
266 |
+
"vision_tower.transformer.layers.7.feed_forward.down_proj",
|
267 |
+
"vision_tower.transformer.layers.7.attention.k_proj",
|
268 |
+
"vision_tower.transformer.layers.7.attention.v_proj",
|
269 |
+
"vision_tower.transformer.layers.7.attention.q_proj",
|
270 |
+
"vision_tower.transformer.layers.7.attention.o_proj",
|
271 |
+
"vision_tower.transformer.layers.8.feed_forward.gate_proj",
|
272 |
+
"vision_tower.transformer.layers.8.feed_forward.up_proj",
|
273 |
+
"vision_tower.transformer.layers.8.feed_forward.down_proj",
|
274 |
+
"vision_tower.transformer.layers.8.attention.k_proj",
|
275 |
+
"vision_tower.transformer.layers.8.attention.v_proj",
|
276 |
+
"vision_tower.transformer.layers.8.attention.q_proj",
|
277 |
+
"vision_tower.transformer.layers.8.attention.o_proj",
|
278 |
+
"vision_tower.transformer.layers.9.feed_forward.gate_proj",
|
279 |
+
"vision_tower.transformer.layers.9.feed_forward.up_proj",
|
280 |
+
"vision_tower.transformer.layers.9.feed_forward.down_proj",
|
281 |
+
"vision_tower.transformer.layers.9.attention.k_proj",
|
282 |
+
"vision_tower.transformer.layers.9.attention.v_proj",
|
283 |
+
"vision_tower.transformer.layers.9.attention.q_proj",
|
284 |
+
"vision_tower.transformer.layers.9.attention.o_proj",
|
285 |
+
"vision_tower.transformer.layers.10.feed_forward.gate_proj",
|
286 |
+
"vision_tower.transformer.layers.10.feed_forward.up_proj",
|
287 |
+
"vision_tower.transformer.layers.10.feed_forward.down_proj",
|
288 |
+
"vision_tower.transformer.layers.10.attention.k_proj",
|
289 |
+
"vision_tower.transformer.layers.10.attention.v_proj",
|
290 |
+
"vision_tower.transformer.layers.10.attention.q_proj",
|
291 |
+
"vision_tower.transformer.layers.10.attention.o_proj",
|
292 |
+
"vision_tower.transformer.layers.11.feed_forward.gate_proj",
|
293 |
+
"vision_tower.transformer.layers.11.feed_forward.up_proj",
|
294 |
+
"vision_tower.transformer.layers.11.feed_forward.down_proj",
|
295 |
+
"vision_tower.transformer.layers.11.attention.k_proj",
|
296 |
+
"vision_tower.transformer.layers.11.attention.v_proj",
|
297 |
+
"vision_tower.transformer.layers.11.attention.q_proj",
|
298 |
+
"vision_tower.transformer.layers.11.attention.o_proj",
|
299 |
+
"vision_tower.transformer.layers.12.feed_forward.gate_proj",
|
300 |
+
"vision_tower.transformer.layers.12.feed_forward.up_proj",
|
301 |
+
"vision_tower.transformer.layers.12.feed_forward.down_proj",
|
302 |
+
"vision_tower.transformer.layers.12.attention.k_proj",
|
303 |
+
"vision_tower.transformer.layers.12.attention.v_proj",
|
304 |
+
"vision_tower.transformer.layers.12.attention.q_proj",
|
305 |
+
"vision_tower.transformer.layers.12.attention.o_proj",
|
306 |
+
"vision_tower.transformer.layers.13.feed_forward.gate_proj",
|
307 |
+
"vision_tower.transformer.layers.13.feed_forward.up_proj",
|
308 |
+
"vision_tower.transformer.layers.13.feed_forward.down_proj",
|
309 |
+
"vision_tower.transformer.layers.13.attention.k_proj",
|
310 |
+
"vision_tower.transformer.layers.13.attention.v_proj",
|
311 |
+
"vision_tower.transformer.layers.13.attention.q_proj",
|
312 |
+
"vision_tower.transformer.layers.13.attention.o_proj",
|
313 |
+
"vision_tower.transformer.layers.14.feed_forward.gate_proj",
|
314 |
+
"vision_tower.transformer.layers.14.feed_forward.up_proj",
|
315 |
+
"vision_tower.transformer.layers.14.feed_forward.down_proj",
|
316 |
+
"vision_tower.transformer.layers.14.attention.k_proj",
|
317 |
+
"vision_tower.transformer.layers.14.attention.v_proj",
|
318 |
+
"vision_tower.transformer.layers.14.attention.q_proj",
|
319 |
+
"vision_tower.transformer.layers.14.attention.o_proj",
|
320 |
+
"vision_tower.transformer.layers.15.feed_forward.gate_proj",
|
321 |
+
"vision_tower.transformer.layers.15.feed_forward.up_proj",
|
322 |
+
"vision_tower.transformer.layers.15.feed_forward.down_proj",
|
323 |
+
"vision_tower.transformer.layers.15.attention.k_proj",
|
324 |
+
"vision_tower.transformer.layers.15.attention.v_proj",
|
325 |
+
"vision_tower.transformer.layers.15.attention.q_proj",
|
326 |
+
"vision_tower.transformer.layers.15.attention.o_proj",
|
327 |
+
"vision_tower.transformer.layers.16.feed_forward.gate_proj",
|
328 |
+
"vision_tower.transformer.layers.16.feed_forward.up_proj",
|
329 |
+
"vision_tower.transformer.layers.16.feed_forward.down_proj",
|
330 |
+
"vision_tower.transformer.layers.16.attention.k_proj",
|
331 |
+
"vision_tower.transformer.layers.16.attention.v_proj",
|
332 |
+
"vision_tower.transformer.layers.16.attention.q_proj",
|
333 |
+
"vision_tower.transformer.layers.16.attention.o_proj",
|
334 |
+
"vision_tower.transformer.layers.17.feed_forward.gate_proj",
|
335 |
+
"vision_tower.transformer.layers.17.feed_forward.up_proj",
|
336 |
+
"vision_tower.transformer.layers.17.feed_forward.down_proj",
|
337 |
+
"vision_tower.transformer.layers.17.attention.k_proj",
|
338 |
+
"vision_tower.transformer.layers.17.attention.v_proj",
|
339 |
+
"vision_tower.transformer.layers.17.attention.q_proj",
|
340 |
+
"vision_tower.transformer.layers.17.attention.o_proj",
|
341 |
+
"vision_tower.transformer.layers.18.feed_forward.gate_proj",
|
342 |
+
"vision_tower.transformer.layers.18.feed_forward.up_proj",
|
343 |
+
"vision_tower.transformer.layers.18.feed_forward.down_proj",
|
344 |
+
"vision_tower.transformer.layers.18.attention.k_proj",
|
345 |
+
"vision_tower.transformer.layers.18.attention.v_proj",
|
346 |
+
"vision_tower.transformer.layers.18.attention.q_proj",
|
347 |
+
"vision_tower.transformer.layers.18.attention.o_proj",
|
348 |
+
"vision_tower.transformer.layers.19.feed_forward.gate_proj",
|
349 |
+
"vision_tower.transformer.layers.19.feed_forward.up_proj",
|
350 |
+
"vision_tower.transformer.layers.19.feed_forward.down_proj",
|
351 |
+
"vision_tower.transformer.layers.19.attention.k_proj",
|
352 |
+
"vision_tower.transformer.layers.19.attention.v_proj",
|
353 |
+
"vision_tower.transformer.layers.19.attention.q_proj",
|
354 |
+
"vision_tower.transformer.layers.19.attention.o_proj",
|
355 |
+
"vision_tower.transformer.layers.20.feed_forward.gate_proj",
|
356 |
+
"vision_tower.transformer.layers.20.feed_forward.up_proj",
|
357 |
+
"vision_tower.transformer.layers.20.feed_forward.down_proj",
|
358 |
+
"vision_tower.transformer.layers.20.attention.k_proj",
|
359 |
+
"vision_tower.transformer.layers.20.attention.v_proj",
|
360 |
+
"vision_tower.transformer.layers.20.attention.q_proj",
|
361 |
+
"vision_tower.transformer.layers.20.attention.o_proj",
|
362 |
+
"vision_tower.transformer.layers.21.feed_forward.gate_proj",
|
363 |
+
"vision_tower.transformer.layers.21.feed_forward.up_proj",
|
364 |
+
"vision_tower.transformer.layers.21.feed_forward.down_proj",
|
365 |
+
"vision_tower.transformer.layers.21.attention.k_proj",
|
366 |
+
"vision_tower.transformer.layers.21.attention.v_proj",
|
367 |
+
"vision_tower.transformer.layers.21.attention.q_proj",
|
368 |
+
"vision_tower.transformer.layers.21.attention.o_proj",
|
369 |
+
"vision_tower.transformer.layers.22.feed_forward.gate_proj",
|
370 |
+
"vision_tower.transformer.layers.22.feed_forward.up_proj",
|
371 |
+
"vision_tower.transformer.layers.22.feed_forward.down_proj",
|
372 |
+
"vision_tower.transformer.layers.22.attention.k_proj",
|
373 |
+
"vision_tower.transformer.layers.22.attention.v_proj",
|
374 |
+
"vision_tower.transformer.layers.22.attention.q_proj",
|
375 |
+
"vision_tower.transformer.layers.22.attention.o_proj",
|
376 |
+
"vision_tower.transformer.layers.23.feed_forward.gate_proj",
|
377 |
+
"vision_tower.transformer.layers.23.feed_forward.up_proj",
|
378 |
+
"vision_tower.transformer.layers.23.feed_forward.down_proj",
|
379 |
+
"vision_tower.transformer.layers.23.attention.k_proj",
|
380 |
+
"vision_tower.transformer.layers.23.attention.v_proj",
|
381 |
+
"vision_tower.transformer.layers.23.attention.q_proj",
|
382 |
+
"vision_tower.transformer.layers.23.attention.o_proj",
|
383 |
+
"multi_modal_projector.linear_1",
|
384 |
+
"multi_modal_projector.linear_2",
|
385 |
+
"language_model.model.layers.0.self_attn.v_proj",
|
386 |
+
"language_model.model.layers.0.self_attn.o_proj",
|
387 |
+
"language_model.model.layers.0.mlp.gate_proj",
|
388 |
+
"language_model.model.layers.0.mlp.up_proj",
|
389 |
+
"language_model.model.layers.0.mlp.down_proj",
|
390 |
+
"language_model.model.layers.1.self_attn.q_proj",
|
391 |
+
"language_model.model.layers.1.self_attn.k_proj",
|
392 |
+
"language_model.model.layers.1.self_attn.v_proj",
|
393 |
+
"language_model.model.layers.1.self_attn.o_proj",
|
394 |
+
"language_model.model.layers.1.mlp.gate_proj",
|
395 |
+
"language_model.model.layers.1.mlp.up_proj",
|
396 |
+
"language_model.model.layers.1.mlp.down_proj",
|
397 |
+
"language_model.model.layers.2.self_attn.q_proj",
|
398 |
+
"language_model.model.layers.2.self_attn.k_proj",
|
399 |
+
"language_model.model.layers.2.self_attn.v_proj",
|
400 |
+
"language_model.model.layers.2.self_attn.o_proj",
|
401 |
+
"language_model.model.layers.2.mlp.gate_proj",
|
402 |
+
"language_model.model.layers.2.mlp.up_proj",
|
403 |
+
"language_model.model.layers.2.mlp.down_proj",
|
404 |
+
"language_model.model.layers.3.self_attn.q_proj",
|
405 |
+
"language_model.model.layers.3.self_attn.k_proj",
|
406 |
+
"language_model.model.layers.3.self_attn.v_proj",
|
407 |
+
"language_model.model.layers.3.self_attn.o_proj",
|
408 |
+
"language_model.model.layers.3.mlp.gate_proj",
|
409 |
+
"language_model.model.layers.3.mlp.up_proj",
|
410 |
+
"language_model.model.layers.3.mlp.down_proj",
|
411 |
+
"language_model.model.layers.4.self_attn.q_proj",
|
412 |
+
"language_model.model.layers.4.self_attn.k_proj",
|
413 |
+
"language_model.model.layers.4.self_attn.v_proj",
|
414 |
+
"language_model.model.layers.4.self_attn.o_proj",
|
415 |
+
"language_model.model.layers.4.mlp.gate_proj",
|
416 |
+
"language_model.model.layers.4.mlp.up_proj",
|
417 |
+
"language_model.model.layers.4.mlp.down_proj",
|
418 |
+
"language_model.model.layers.5.self_attn.q_proj",
|
419 |
+
"language_model.model.layers.5.self_attn.k_proj",
|
420 |
+
"language_model.model.layers.5.self_attn.v_proj",
|
421 |
+
"language_model.model.layers.5.self_attn.o_proj",
|
422 |
+
"language_model.model.layers.5.mlp.gate_proj",
|
423 |
+
"language_model.model.layers.5.mlp.up_proj",
|
424 |
+
"language_model.model.layers.5.mlp.down_proj",
|
425 |
+
"language_model.model.layers.6.self_attn.q_proj",
|
426 |
+
"language_model.model.layers.6.self_attn.k_proj",
|
427 |
+
"language_model.model.layers.6.self_attn.v_proj",
|
428 |
+
"language_model.model.layers.6.self_attn.o_proj",
|
429 |
+
"language_model.model.layers.6.mlp.gate_proj",
|
430 |
+
"language_model.model.layers.6.mlp.up_proj",
|
431 |
+
"language_model.model.layers.6.mlp.down_proj",
|
432 |
+
"language_model.model.layers.7.self_attn.q_proj",
|
433 |
+
"language_model.model.layers.7.self_attn.k_proj",
|
434 |
+
"language_model.model.layers.7.self_attn.v_proj",
|
435 |
+
"language_model.model.layers.7.self_attn.o_proj",
|
436 |
+
"language_model.model.layers.7.mlp.gate_proj",
|
437 |
+
"language_model.model.layers.7.mlp.up_proj",
|
438 |
+
"language_model.model.layers.7.mlp.down_proj",
|
439 |
+
"language_model.model.layers.8.self_attn.q_proj",
|
440 |
+
"language_model.model.layers.8.self_attn.k_proj",
|
441 |
+
"language_model.model.layers.8.self_attn.v_proj",
|
442 |
+
"language_model.model.layers.8.self_attn.o_proj",
|
443 |
+
"language_model.model.layers.8.mlp.gate_proj",
|
444 |
+
"language_model.model.layers.8.mlp.up_proj",
|
445 |
+
"language_model.model.layers.8.mlp.down_proj",
|
446 |
+
"language_model.model.layers.9.self_attn.q_proj",
|
447 |
+
"language_model.model.layers.9.self_attn.k_proj",
|
448 |
+
"language_model.model.layers.9.self_attn.v_proj",
|
449 |
+
"language_model.model.layers.9.self_attn.o_proj",
|
450 |
+
"language_model.model.layers.9.mlp.gate_proj",
|
451 |
+
"language_model.model.layers.9.mlp.up_proj",
|
452 |
+
"language_model.model.layers.9.mlp.down_proj",
|
453 |
+
"language_model.model.layers.10.self_attn.q_proj",
|
454 |
+
"language_model.model.layers.10.self_attn.k_proj",
|
455 |
+
"language_model.model.layers.10.self_attn.v_proj",
|
456 |
+
"language_model.model.layers.10.self_attn.o_proj",
|
457 |
+
"language_model.model.layers.10.mlp.gate_proj",
|
458 |
+
"language_model.model.layers.10.mlp.up_proj",
|
459 |
+
"language_model.model.layers.10.mlp.down_proj",
|
460 |
+
"language_model.model.layers.11.self_attn.q_proj",
|
461 |
+
"language_model.model.layers.11.self_attn.k_proj",
|
462 |
+
"language_model.model.layers.11.self_attn.v_proj",
|
463 |
+
"language_model.model.layers.11.self_attn.o_proj",
|
464 |
+
"language_model.model.layers.11.mlp.gate_proj",
|
465 |
+
"language_model.model.layers.11.mlp.up_proj",
|
466 |
+
"language_model.model.layers.11.mlp.down_proj",
|
467 |
+
"language_model.model.layers.12.self_attn.q_proj",
|
468 |
+
"language_model.model.layers.12.self_attn.k_proj",
|
469 |
+
"language_model.model.layers.12.self_attn.v_proj",
|
470 |
+
"language_model.model.layers.12.self_attn.o_proj",
|
471 |
+
"language_model.model.layers.12.mlp.gate_proj",
|
472 |
+
"language_model.model.layers.12.mlp.up_proj",
|
473 |
+
"language_model.model.layers.12.mlp.down_proj",
|
474 |
+
"language_model.model.layers.13.self_attn.q_proj",
|
475 |
+
"language_model.model.layers.13.self_attn.k_proj",
|
476 |
+
"language_model.model.layers.13.self_attn.v_proj",
|
477 |
+
"language_model.model.layers.13.self_attn.o_proj",
|
478 |
+
"language_model.model.layers.13.mlp.gate_proj",
|
479 |
+
"language_model.model.layers.13.mlp.up_proj",
|
480 |
+
"language_model.model.layers.13.mlp.down_proj",
|
481 |
+
"language_model.model.layers.14.self_attn.q_proj",
|
482 |
+
"language_model.model.layers.14.self_attn.k_proj",
|
483 |
+
"language_model.model.layers.14.self_attn.v_proj",
|
484 |
+
"language_model.model.layers.14.self_attn.o_proj",
|
485 |
+
"language_model.model.layers.14.mlp.gate_proj",
|
486 |
+
"language_model.model.layers.14.mlp.up_proj",
|
487 |
+
"language_model.model.layers.14.mlp.down_proj",
|
488 |
+
"language_model.model.layers.15.self_attn.q_proj",
|
489 |
+
"language_model.model.layers.15.self_attn.k_proj",
|
490 |
+
"language_model.model.layers.15.self_attn.v_proj",
|
491 |
+
"language_model.model.layers.15.self_attn.o_proj",
|
492 |
+
"language_model.model.layers.15.mlp.gate_proj",
|
493 |
+
"language_model.model.layers.15.mlp.up_proj",
|
494 |
+
"language_model.model.layers.15.mlp.down_proj",
|
495 |
+
"language_model.model.layers.16.self_attn.q_proj",
|
496 |
+
"language_model.model.layers.16.self_attn.k_proj",
|
497 |
+
"language_model.model.layers.16.self_attn.v_proj",
|
498 |
+
"language_model.model.layers.16.self_attn.o_proj",
|
499 |
+
"language_model.model.layers.16.mlp.gate_proj",
|
500 |
+
"language_model.model.layers.16.mlp.up_proj",
|
501 |
+
"language_model.model.layers.16.mlp.down_proj",
|
502 |
+
"language_model.model.layers.17.self_attn.q_proj",
|
503 |
+
"language_model.model.layers.17.self_attn.k_proj",
|
504 |
+
"language_model.model.layers.17.self_attn.v_proj",
|
505 |
+
"language_model.model.layers.17.self_attn.o_proj",
|
506 |
+
"language_model.model.layers.17.mlp.gate_proj",
|
507 |
+
"language_model.model.layers.17.mlp.up_proj",
|
508 |
+
"language_model.model.layers.17.mlp.down_proj",
|
509 |
+
"language_model.model.layers.18.self_attn.q_proj",
|
510 |
+
"language_model.model.layers.18.self_attn.k_proj",
|
511 |
+
"language_model.model.layers.18.self_attn.v_proj",
|
512 |
+
"language_model.model.layers.18.self_attn.o_proj",
|
513 |
+
"language_model.model.layers.18.mlp.gate_proj",
|
514 |
+
"language_model.model.layers.18.mlp.up_proj",
|
515 |
+
"language_model.model.layers.18.mlp.down_proj",
|
516 |
+
"language_model.model.layers.19.self_attn.q_proj",
|
517 |
+
"language_model.model.layers.19.self_attn.k_proj",
|
518 |
+
"language_model.model.layers.19.self_attn.v_proj",
|
519 |
+
"language_model.model.layers.19.self_attn.o_proj",
|
520 |
+
"language_model.model.layers.19.mlp.gate_proj",
|
521 |
+
"language_model.model.layers.19.mlp.up_proj",
|
522 |
+
"language_model.model.layers.19.mlp.down_proj",
|
523 |
+
"language_model.model.layers.20.self_attn.q_proj",
|
524 |
+
"language_model.model.layers.20.self_attn.k_proj",
|
525 |
+
"language_model.model.layers.20.self_attn.v_proj",
|
526 |
+
"language_model.model.layers.20.self_attn.o_proj",
|
527 |
+
"language_model.model.layers.20.mlp.gate_proj",
|
528 |
+
"language_model.model.layers.20.mlp.up_proj",
|
529 |
+
"language_model.model.layers.20.mlp.down_proj",
|
530 |
+
"language_model.model.layers.21.self_attn.q_proj",
|
531 |
+
"language_model.model.layers.21.self_attn.k_proj",
|
532 |
+
"language_model.model.layers.21.self_attn.v_proj",
|
533 |
+
"language_model.model.layers.21.self_attn.o_proj",
|
534 |
+
"language_model.model.layers.21.mlp.gate_proj",
|
535 |
+
"language_model.model.layers.21.mlp.up_proj",
|
536 |
+
"language_model.model.layers.21.mlp.down_proj",
|
537 |
+
"language_model.model.layers.22.self_attn.q_proj",
|
538 |
+
"language_model.model.layers.22.self_attn.k_proj",
|
539 |
+
"language_model.model.layers.22.self_attn.v_proj",
|
540 |
+
"language_model.model.layers.22.self_attn.o_proj",
|
541 |
+
"language_model.model.layers.22.mlp.gate_proj",
|
542 |
+
"language_model.model.layers.22.mlp.up_proj",
|
543 |
+
"language_model.model.layers.22.mlp.down_proj",
|
544 |
+
"language_model.model.layers.23.self_attn.q_proj",
|
545 |
+
"language_model.model.layers.23.self_attn.k_proj",
|
546 |
+
"language_model.model.layers.23.self_attn.v_proj",
|
547 |
+
"language_model.model.layers.23.self_attn.o_proj",
|
548 |
+
"language_model.model.layers.23.mlp.gate_proj",
|
549 |
+
"language_model.model.layers.23.mlp.up_proj",
|
550 |
+
"language_model.model.layers.23.mlp.down_proj",
|
551 |
+
"language_model.model.layers.24.self_attn.q_proj",
|
552 |
+
"language_model.model.layers.24.self_attn.k_proj",
|
553 |
+
"language_model.model.layers.24.self_attn.v_proj",
|
554 |
+
"language_model.model.layers.24.self_attn.o_proj",
|
555 |
+
"language_model.model.layers.24.mlp.gate_proj",
|
556 |
+
"language_model.model.layers.24.mlp.up_proj",
|
557 |
+
"language_model.model.layers.24.mlp.down_proj",
|
558 |
+
"language_model.model.layers.25.self_attn.q_proj",
|
559 |
+
"language_model.model.layers.25.self_attn.k_proj",
|
560 |
+
"language_model.model.layers.25.self_attn.v_proj",
|
561 |
+
"language_model.model.layers.25.self_attn.o_proj",
|
562 |
+
"language_model.model.layers.25.mlp.gate_proj",
|
563 |
+
"language_model.model.layers.25.mlp.up_proj",
|
564 |
+
"language_model.model.layers.25.mlp.down_proj",
|
565 |
+
"language_model.model.layers.26.self_attn.q_proj",
|
566 |
+
"language_model.model.layers.26.self_attn.k_proj",
|
567 |
+
"language_model.model.layers.26.self_attn.v_proj",
|
568 |
+
"language_model.model.layers.26.self_attn.o_proj",
|
569 |
+
"language_model.model.layers.26.mlp.gate_proj",
|
570 |
+
"language_model.model.layers.26.mlp.up_proj",
|
571 |
+
"language_model.model.layers.26.mlp.down_proj",
|
572 |
+
"language_model.model.layers.27.self_attn.q_proj",
|
573 |
+
"language_model.model.layers.27.self_attn.k_proj",
|
574 |
+
"language_model.model.layers.27.self_attn.v_proj",
|
575 |
+
"language_model.model.layers.27.self_attn.o_proj",
|
576 |
+
"language_model.model.layers.27.mlp.gate_proj",
|
577 |
+
"language_model.model.layers.27.mlp.up_proj",
|
578 |
+
"language_model.model.layers.27.mlp.down_proj",
|
579 |
+
"language_model.model.layers.28.self_attn.q_proj",
|
580 |
+
"language_model.model.layers.28.self_attn.k_proj",
|
581 |
+
"language_model.model.layers.28.self_attn.v_proj",
|
582 |
+
"language_model.model.layers.28.self_attn.o_proj",
|
583 |
+
"language_model.model.layers.28.mlp.gate_proj",
|
584 |
+
"language_model.model.layers.28.mlp.up_proj",
|
585 |
+
"language_model.model.layers.28.mlp.down_proj",
|
586 |
+
"language_model.model.layers.29.self_attn.q_proj",
|
587 |
+
"language_model.model.layers.29.self_attn.k_proj",
|
588 |
+
"language_model.model.layers.29.self_attn.v_proj",
|
589 |
+
"language_model.model.layers.29.self_attn.o_proj",
|
590 |
+
"language_model.model.layers.29.mlp.gate_proj",
|
591 |
+
"language_model.model.layers.29.mlp.up_proj",
|
592 |
+
"language_model.model.layers.29.mlp.down_proj",
|
593 |
+
"language_model.model.layers.30.self_attn.q_proj",
|
594 |
+
"language_model.model.layers.30.self_attn.k_proj",
|
595 |
+
"language_model.model.layers.30.self_attn.v_proj",
|
596 |
+
"language_model.model.layers.30.self_attn.o_proj",
|
597 |
+
"language_model.model.layers.30.mlp.gate_proj",
|
598 |
+
"language_model.model.layers.30.mlp.up_proj",
|
599 |
+
"language_model.model.layers.30.mlp.down_proj",
|
600 |
+
"language_model.model.layers.31.self_attn.q_proj",
|
601 |
+
"language_model.model.layers.31.self_attn.k_proj",
|
602 |
+
"language_model.model.layers.31.self_attn.v_proj",
|
603 |
+
"language_model.model.layers.31.self_attn.o_proj",
|
604 |
+
"language_model.model.layers.31.mlp.gate_proj",
|
605 |
+
"language_model.model.layers.31.mlp.up_proj",
|
606 |
+
"language_model.model.layers.31.mlp.down_proj",
|
607 |
+
"language_model.model.layers.32.self_attn.q_proj",
|
608 |
+
"language_model.model.layers.32.self_attn.k_proj",
|
609 |
+
"language_model.model.layers.32.self_attn.v_proj",
|
610 |
+
"language_model.model.layers.32.self_attn.o_proj",
|
611 |
+
"language_model.model.layers.32.mlp.gate_proj",
|
612 |
+
"language_model.model.layers.32.mlp.up_proj",
|
613 |
+
"language_model.model.layers.32.mlp.down_proj",
|
614 |
+
"language_model.model.layers.33.self_attn.q_proj",
|
615 |
+
"language_model.model.layers.33.self_attn.k_proj",
|
616 |
+
"language_model.model.layers.33.self_attn.v_proj",
|
617 |
+
"language_model.model.layers.33.self_attn.o_proj",
|
618 |
+
"language_model.model.layers.33.mlp.gate_proj",
|
619 |
+
"language_model.model.layers.33.mlp.up_proj",
|
620 |
+
"language_model.model.layers.33.mlp.down_proj",
|
621 |
+
"language_model.model.layers.34.self_attn.q_proj",
|
622 |
+
"language_model.model.layers.34.self_attn.k_proj",
|
623 |
+
"language_model.model.layers.34.self_attn.v_proj",
|
624 |
+
"language_model.model.layers.34.self_attn.o_proj",
|
625 |
+
"language_model.model.layers.34.mlp.gate_proj",
|
626 |
+
"language_model.model.layers.34.mlp.up_proj",
|
627 |
+
"language_model.model.layers.34.mlp.down_proj",
|
628 |
+
"language_model.model.layers.35.self_attn.q_proj",
|
629 |
+
"language_model.model.layers.35.self_attn.k_proj",
|
630 |
+
"language_model.model.layers.35.self_attn.v_proj",
|
631 |
+
"language_model.model.layers.35.self_attn.o_proj",
|
632 |
+
"language_model.model.layers.35.mlp.gate_proj",
|
633 |
+
"language_model.model.layers.35.mlp.up_proj",
|
634 |
+
"language_model.model.layers.35.mlp.down_proj",
|
635 |
+
"language_model.model.layers.36.self_attn.q_proj",
|
636 |
+
"language_model.model.layers.36.self_attn.k_proj",
|
637 |
+
"language_model.model.layers.36.self_attn.v_proj",
|
638 |
+
"language_model.model.layers.36.self_attn.o_proj",
|
639 |
+
"language_model.model.layers.36.mlp.gate_proj",
|
640 |
+
"language_model.model.layers.36.mlp.up_proj",
|
641 |
+
"language_model.model.layers.36.mlp.down_proj",
|
642 |
+
"language_model.model.layers.37.self_attn.q_proj",
|
643 |
+
"language_model.model.layers.37.self_attn.k_proj",
|
644 |
+
"language_model.model.layers.37.self_attn.v_proj",
|
645 |
+
"language_model.model.layers.37.self_attn.o_proj",
|
646 |
+
"language_model.model.layers.37.mlp.gate_proj",
|
647 |
+
"language_model.model.layers.37.mlp.up_proj",
|
648 |
+
"language_model.model.layers.37.mlp.down_proj",
|
649 |
+
"language_model.model.layers.38.self_attn.q_proj",
|
650 |
+
"language_model.model.layers.38.self_attn.k_proj",
|
651 |
+
"language_model.model.layers.38.self_attn.v_proj",
|
652 |
+
"language_model.model.layers.38.self_attn.o_proj",
|
653 |
+
"language_model.model.layers.38.mlp.gate_proj",
|
654 |
+
"language_model.model.layers.38.mlp.up_proj",
|
655 |
+
"language_model.model.layers.38.mlp.down_proj",
|
656 |
+
"language_model.model.layers.39.self_attn.q_proj",
|
657 |
+
"language_model.model.layers.39.self_attn.k_proj",
|
658 |
+
"language_model.model.layers.39.self_attn.v_proj",
|
659 |
+
"language_model.model.layers.39.self_attn.o_proj",
|
660 |
+
"language_model.model.layers.39.mlp.gate_proj",
|
661 |
+
"language_model.model.layers.39.mlp.up_proj",
|
662 |
+
"language_model.model.layers.39.mlp.down_proj",
|
663 |
+
"language_model.lm_head"
|
664 |
+
],
|
665 |
+
"registry_requires_subclass": false,
|
666 |
+
"sparsity_structure": "unstructured",
|
667 |
+
"targets": [
|
668 |
+
"language_model.model.layers.0.self_attn.q_proj",
|
669 |
+
"language_model.model.layers.0.self_attn.k_proj"
|
670 |
+
]
|
671 |
+
}
|
672 |
+
},
|
673 |
+
"text_config": {
|
674 |
+
"hidden_size": 5120,
|
675 |
+
"intermediate_size": 14336,
|
676 |
+
"is_composition": true,
|
677 |
+
"max_position_embeddings": 1024000,
|
678 |
+
"model_type": "mistral",
|
679 |
+
"num_hidden_layers": 40,
|
680 |
+
"num_key_value_heads": 8,
|
681 |
+
"rms_norm_eps": 1e-05,
|
682 |
+
"rope_theta": 1000000000.0,
|
683 |
+
"sliding_window": null,
|
684 |
+
"vocab_size": 131072
|
685 |
+
},
|
686 |
+
"torch_dtype": "bfloat16",
|
687 |
+
"transformers_version": "4.47.0",
|
688 |
+
"vision_config": {
|
689 |
+
"head_dim": 64,
|
690 |
+
"hidden_act": "silu",
|
691 |
+
"image_size": 1024,
|
692 |
+
"is_composition": true,
|
693 |
+
"model_type": "pixtral",
|
694 |
+
"patch_size": 16,
|
695 |
+
"rope_theta": 10000.0,
|
696 |
+
"tie_word_embeddings": false
|
697 |
+
},
|
698 |
+
"vision_feature_layer": -1,
|
699 |
+
"vision_feature_select_strategy": "full"
|
700 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 1,
|
4 |
+
"eos_token_id": 2,
|
5 |
+
"transformers_version": "4.47.0"
|
6 |
+
}
|
model-00001-of-00002.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:203b340ed25f81ab50e7a98dd061a1dea966c96302c1d41a89e2c18b55d19cb1
|
3 |
+
size 4986219584
|
model-00002-of-00002.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:02865146fab29753c6c9208cb108e845679d8c164d8fbd92be657a05625a673a
|
3 |
+
size 4192021568
|
model.safetensors.index.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
preprocessor_config.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_convert_rgb": true,
|
3 |
+
"do_normalize": true,
|
4 |
+
"do_rescale": true,
|
5 |
+
"do_resize": true,
|
6 |
+
"image_mean": [
|
7 |
+
0.48145466,
|
8 |
+
0.4578275,
|
9 |
+
0.40821073
|
10 |
+
],
|
11 |
+
"image_processor_type": "PixtralImageProcessorFast",
|
12 |
+
"image_std": [
|
13 |
+
0.26862954,
|
14 |
+
0.26130258,
|
15 |
+
0.27577711
|
16 |
+
],
|
17 |
+
"patch_size": {
|
18 |
+
"height": 16,
|
19 |
+
"width": 16
|
20 |
+
},
|
21 |
+
"processor_class": "PixtralProcessor",
|
22 |
+
"resample": 3,
|
23 |
+
"rescale_factor": 0.00392156862745098,
|
24 |
+
"size": {
|
25 |
+
"longest_edge": 1024
|
26 |
+
}
|
27 |
+
}
|
processor_config.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"image_break_token": "[IMG_BREAK]",
|
3 |
+
"image_end_token": "[IMG_END]",
|
4 |
+
"image_token": "[IMG]",
|
5 |
+
"patch_size": 16,
|
6 |
+
"processor_class": "PixtralProcessor"
|
7 |
+
}
|
recipe.yaml
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
DEFAULT_stage:
|
2 |
+
DEFAULT_modifiers:
|
3 |
+
GPTQModifier:
|
4 |
+
dampening_frac: 0.05
|
5 |
+
config_groups:
|
6 |
+
config_group:
|
7 |
+
targets: [Linear]
|
8 |
+
weights: {num_bits: 4, type: int, symmetric: true, group_size: 128, strategy: group,
|
9 |
+
dynamic: false, actorder: weight, observer: minmax}
|
10 |
+
targets: Linear
|
11 |
+
ignore: ['re:.*lm_head', 're:multi_modal_projector.*', 're:vision_tower.*']
|
special_tokens_map.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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": "</s>",
|
17 |
+
"unk_token": {
|
18 |
+
"content": "<unk>",
|
19 |
+
"lstrip": false,
|
20 |
+
"normalized": false,
|
21 |
+
"rstrip": false,
|
22 |
+
"single_word": false
|
23 |
+
}
|
24 |
+
}
|
tokenizer.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fcdf3e6b96371b9a5c8673bdf2f1bc838b994b73a6fa995dca62196346aaddb5
|
3 |
+
size 17077311
|
tokenizer_config.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|