Hugging Face's logo

ScienceOne-AI
/
S1-Omni

Safetensors
qwen3_vl
Model card Files
xet
 Community
2
ScienceOne-AI commited on
Commit
835604b
·
verified ·
1 Parent(s): cde174b

Upload folder using huggingface_hub

Browse files
Files changed (23) hide show
  1. .gitattributes +1 -0
  2. .ms_upload_cache +1 -0
  3. added_tokens.json +35 -0
  4. chat_template.jinja +110 -0
  5. config.json +479 -0
  6. configuration.json +1 -0
  7. esm2/config.json +49 -0
  8. esm2/esm2_config.json +49 -0
  9. esm2/tokenizer_config.json +54 -0
  10. esm2/vocab.txt +33 -0
  11. generation_config.json +13 -0
  12. model.safetensors.index.json +0 -0
  13. preprocessor_config.json +21 -0
  14. processor_config.json +62 -0
  15. s1_omni_image_config.json +208 -0
  16. s1_protein_config.json +17 -0
  17. simplefold.safetensors.index.json +418 -0
  18. simplefold_config.json +116 -0
  19. special_tokens_map.json +38 -0
  20. tokenizer.json +3 -0
  21. tokenizer_config.json +303 -0
  22. video_preprocessor_config.json +41 -0
  23. vocab.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
.ms_upload_cache ADDED
@@ -0,0 +1 @@
 
 
1
+ {"version": 3, "repo_id": "ScienceOne-AI/S1-Omni", "files": {"esm2/tokenizer_config.json|1784107455.719805|1158": {"hash": "8f0ffecfd354841b60ce5266a9b23686a614215e02c433aa1ba048798514bc06", "size": 1158}, "esm2/config.json|1784109184.1885962|1158": {"hash": "64ceb18d10a221b07b2e6fe799f8f4e6f198721adc8941a9edc9d033f90fe5df", "size": 1158}, "added_tokens.json|1784109617.8940053|883": {"hash": "c79800872e3618518d48485c291a7382ca1b54bde005854756c3e2d999fb6344", "size": 883, "status": "c"}, "configuration.json|1784107374.6217923|51": {"hash": "2d4464e2ead06bc9bc718c781309ad1e7baded626d66e8dcdc8b469ba185faf0", "size": 51, "status": "c"}, "esm2/esm2_config.json|1784109216.107153|1158": {"hash": "64ceb18d10a221b07b2e6fe799f8f4e6f198721adc8941a9edc9d033f90fe5df", "size": 1158}, "chat_template.jinja|1784107369.119209|5199": {"hash": "36e042fe45641f067b1f2381fcc8955d10d956a3ed333ecdf7f7eb0916f68956", "size": 5199, "status": "c"}, "config.json|1784109287.58904|11970": {"hash": "41e31b472022e082e77dbe29bc5af53e9933b0e95c6dfb11727896bf3ca52bed", "size": 11970, "status": "c"}, "esm2/vocab.txt|1784107455.7217665|93": {"hash": "0b82cc0a7c7cf9e567b1e5892d793285b9fbae822c964ca48696f7db44598e03", "size": 93}, "generation_config.json|1784107391.7293556|214": {"hash": "6ae7c44dde74623cb7cfd45c1657964e70ab5d791089f0a1c39dc5e6779007cb", "size": 214}}}
added_tokens.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</think>": 151668,
3
+ "</tool_call>": 151658,
4
+ "</tool_response>": 151666,
5
+ "<think>": 151667,
6
+ "<tool_call>": 151657,
7
+ "<tool_response>": 151665,
8
+ "<image_gen>": 151669,
9
+ "<image_edit>": 151670,
10
+ "<linear_cla>": 151671,
11
+ "<linear_pre>": 151672,
12
+ "<prot_st>": 151673,
13
+ "<prot_cla>": 151674,
14
+ "<spectra_st>": 151675,
15
+ "<|box_end|>": 151649,
16
+ "<|box_start|>": 151648,
17
+ "<|endoftext|>": 151643,
18
+ "<|file_sep|>": 151664,
19
+ "<|fim_middle|>": 151660,
20
+ "<|fim_pad|>": 151662,
21
+ "<|fim_prefix|>": 151659,
22
+ "<|fim_suffix|>": 151661,
23
+ "<|im_end|>": 151645,
24
+ "<|im_start|>": 151644,
25
+ "<|image_pad|>": 151655,
26
+ "<|object_ref_end|>": 151647,
27
+ "<|object_ref_start|>": 151646,
28
+ "<|quad_end|>": 151651,
29
+ "<|quad_start|>": 151650,
30
+ "<|repo_name|>": 151663,
31
+ "<|video_pad|>": 151656,
32
+ "<|vision_end|>": 151653,
33
+ "<|vision_pad|>": 151654,
34
+ "<|vision_start|>": 151652
35
+ }
chat_template.jinja ADDED
@@ -0,0 +1,110 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- set image_count = namespace(value=0) %}
2
+ {%- set video_count = namespace(value=0) %}
3
+ {%- macro render_content(content, do_vision_count) %}
4
+ {%- if content is string %}
5
+ {{- content }}
6
+ {%- else %}
7
+ {%- for item in content %}
8
+ {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
9
+ {%- if do_vision_count %}
10
+ {%- set image_count.value = image_count.value + 1 %}
11
+ {%- endif %}
12
+ {%- if add_vision_id %}Picture {{ image_count.value }}: {% endif -%}
13
+ <|vision_start|><|image_pad|><|vision_end|>
14
+ {%- elif 'video' in item or item.type == 'video' %}
15
+ {%- if do_vision_count %}
16
+ {%- set video_count.value = video_count.value + 1 %}
17
+ {%- endif %}
18
+ {%- if add_vision_id %}Video {{ video_count.value }}: {% endif -%}
19
+ <|vision_start|><|video_pad|><|vision_end|>
20
+ {%- elif 'text' in item %}
21
+ {{- item.text }}
22
+ {%- endif %}
23
+ {%- endfor %}
24
+ {%- endif %}
25
+ {%- endmacro %}
26
+ {%- if tools %}
27
+ {{- '<|im_start|>system\n' }}
28
+ {%- if messages[0].role == 'system' %}
29
+ {{- render_content(messages[0].content, false) + '\n\n' }}
30
+ {%- endif %}
31
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
32
+ {%- for tool in tools %}
33
+ {{- "\n" }}
34
+ {{- tool | tojson }}
35
+ {%- endfor %}
36
+ {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
37
+ {%- else %}
38
+ {%- if messages[0].role == 'system' %}
39
+ {{- '<|im_start|>system\n' + render_content(messages[0].content, false) + '<|im_end|>\n' }}
40
+ {%- endif %}
41
+ {%- endif %}
42
+ {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
43
+ {%- for message in messages[::-1] %}
44
+ {%- set index = (messages|length - 1) - loop.index0 %}
45
+ {%- if ns.multi_step_tool and message.role == "user" %}
46
+ {%- set content = render_content(message.content, false) %}
47
+ {%- if not(content.startswith('<tool_response>') and content.endswith('</tool_response>')) %}
48
+ {%- set ns.multi_step_tool = false %}
49
+ {%- set ns.last_query_index = index %}
50
+ {%- endif %}
51
+ {%- endif %}
52
+ {%- endfor %}
53
+ {%- for message in messages %}
54
+ {%- set content = render_content(message.content, True) %}
55
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
56
+ {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
57
+ {%- elif message.role == "assistant" %}
58
+ {%- set reasoning_content = '' %}
59
+ {%- if message.reasoning_content is string %}
60
+ {%- set reasoning_content = message.reasoning_content %}
61
+ {%- else %}
62
+ {%- if '</think>' in content %}
63
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
64
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
65
+ {%- endif %}
66
+ {%- endif %}
67
+ {%- if loop.index0 > ns.last_query_index %}
68
+ {%- if loop.last or (not loop.last and reasoning_content) %}
69
+ {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
70
+ {%- else %}
71
+ {{- '<|im_start|>' + message.role + '\n' + content }}
72
+ {%- endif %}
73
+ {%- else %}
74
+ {{- '<|im_start|>' + message.role + '\n' + content }}
75
+ {%- endif %}
76
+ {%- if message.tool_calls %}
77
+ {%- for tool_call in message.tool_calls %}
78
+ {%- if (loop.first and content) or (not loop.first) %}
79
+ {{- '\n' }}
80
+ {%- endif %}
81
+ {%- if tool_call.function %}
82
+ {%- set tool_call = tool_call.function %}
83
+ {%- endif %}
84
+ {{- '<tool_call>\n{"name": "' }}
85
+ {{- tool_call.name }}
86
+ {{- '", "arguments": ' }}
87
+ {%- if tool_call.arguments is string %}
88
+ {{- tool_call.arguments }}
89
+ {%- else %}
90
+ {{- tool_call.arguments | tojson }}
91
+ {%- endif %}
92
+ {{- '}\n</tool_call>' }}
93
+ {%- endfor %}
94
+ {%- endif %}
95
+ {{- '<|im_end|>\n' }}
96
+ {%- elif message.role == "tool" %}
97
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
98
+ {{- '<|im_start|>user' }}
99
+ {%- endif %}
100
+ {{- '\n<tool_response>\n' }}
101
+ {{- content }}
102
+ {{- '\n</tool_response>' }}
103
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
104
+ {{- '<|im_end|>\n' }}
105
+ {%- endif %}
106
+ {%- endif %}
107
+ {%- endfor %}
108
+ {%- if add_generation_prompt %}
109
+ {{- '<|im_start|>assistant\n<think>\n' }}
110
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,479 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3VLForConditionalGeneration"
4
+ ],
5
+ "dtype": "bfloat16",
6
+ "eos_token_id": 151645,
7
+ "hidden_size": 5120,
8
+ "image_token_id": 151655,
9
+ "model_type": "qwen3_vl",
10
+ "pad_token_id": 151643,
11
+ "text_config": {
12
+ "attention_bias": false,
13
+ "attention_dropout": 0.0,
14
+ "bos_token_id": 151643,
15
+ "dtype": "bfloat16",
16
+ "eos_token_id": 151645,
17
+ "head_dim": 128,
18
+ "hidden_act": "silu",
19
+ "hidden_size": 5120,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 25600,
22
+ "max_position_embeddings": 262144,
23
+ "model_type": "qwen3_vl_text",
24
+ "num_attention_heads": 64,
25
+ "num_hidden_layers": 64,
26
+ "num_key_value_heads": 8,
27
+ "pad_token_id": 151643,
28
+ "rms_norm_eps": 1e-06,
29
+ "rope_scaling": {
30
+ "mrope_interleaved": true,
31
+ "mrope_section": [
32
+ 24,
33
+ 20,
34
+ 20
35
+ ],
36
+ "rope_type": "default"
37
+ },
38
+ "rope_theta": 5000000,
39
+ "use_cache": false,
40
+ "vocab_size": 151936
41
+ },
42
+ "tie_word_embeddings": false,
43
+ "transformers_version": "4.57.6",
44
+ "video_token_id": 151656,
45
+ "vision_config": {
46
+ "deepstack_visual_indexes": [
47
+ 8,
48
+ 16,
49
+ 24
50
+ ],
51
+ "depth": 27,
52
+ "dtype": "bfloat16",
53
+ "hidden_act": "gelu_pytorch_tanh",
54
+ "hidden_size": 1152,
55
+ "in_channels": 3,
56
+ "initializer_range": 0.02,
57
+ "intermediate_size": 4304,
58
+ "model_type": "qwen3_vl",
59
+ "num_heads": 16,
60
+ "num_position_embeddings": 2304,
61
+ "out_hidden_size": 5120,
62
+ "pad_token_id": 151643,
63
+ "patch_size": 16,
64
+ "spatial_merge_size": 2,
65
+ "temporal_patch_size": 2
66
+ },
67
+ "vision_end_token_id": 151653,
68
+ "vision_start_token_id": 151652,
69
+ "spec2mol_v2": {
70
+ "format": "spec2mol_v2_hf_inline_safetensors",
71
+ "tensor_prefix": "spec2mol.",
72
+ "small_modules": {
73
+ "format": "spec2mol_small_modules_v2",
74
+ "tensor_prefix": "spec2mol.",
75
+ "modules": {
76
+ "s1_omni_ir_linear": {
77
+ "prefix": "s1_omni_ir_linear.",
78
+ "modality": "ir",
79
+ "num_classes": 20,
80
+ "vocab": [
81
+ "alkane",
82
+ "alkene",
83
+ "alkyne",
84
+ "amine",
85
+ "imine",
86
+ "nitrile",
87
+ "alcohol",
88
+ "ether",
89
+ "haloalkane",
90
+ "aldehyde",
91
+ "ketone",
92
+ "ester",
93
+ "amide",
94
+ "arene",
95
+ "imidazole",
96
+ "pyrazole",
97
+ "oxazole",
98
+ "isoxazole",
99
+ "cyclopropane",
100
+ "epoxide"
101
+ ],
102
+ "threshold_source": "0.5",
103
+ "thresholds": [
104
+ 0.5,
105
+ 0.5,
106
+ 0.5,
107
+ 0.5,
108
+ 0.5,
109
+ 0.5,
110
+ 0.5,
111
+ 0.5,
112
+ 0.5,
113
+ 0.5,
114
+ 0.5,
115
+ 0.5,
116
+ 0.5,
117
+ 0.5,
118
+ 0.5,
119
+ 0.5,
120
+ 0.5,
121
+ 0.5,
122
+ 0.5,
123
+ 0.5
124
+ ],
125
+ "build_args": {
126
+ "in_dim": 5120,
127
+ "proj_dim": 768,
128
+ "n_layers": 2,
129
+ "n_heads": 8,
130
+ "ffn_mult": 2,
131
+ "dropout": 0.1,
132
+ "num_queries": 4,
133
+ "num_classes": 20,
134
+ "feat_dim": 0,
135
+ "head_dropout": 0.1
136
+ },
137
+ "hf_prefix": "spec2mol.s1_omni_ir_linear.",
138
+ "tensor_count": 46
139
+ },
140
+ "s1_omni_raman_linear": {
141
+ "prefix": "s1_omni_raman_linear.",
142
+ "modality": "raman",
143
+ "num_classes": 20,
144
+ "vocab": [
145
+ "alkane",
146
+ "alkene",
147
+ "alkyne",
148
+ "amine",
149
+ "imine",
150
+ "nitrile",
151
+ "alcohol",
152
+ "ether",
153
+ "haloalkane",
154
+ "aldehyde",
155
+ "ketone",
156
+ "ester",
157
+ "amide",
158
+ "arene",
159
+ "imidazole",
160
+ "pyrazole",
161
+ "oxazole",
162
+ "isoxazole",
163
+ "cyclopropane",
164
+ "epoxide"
165
+ ],
166
+ "threshold_source": "0.5",
167
+ "thresholds": [
168
+ 0.5,
169
+ 0.5,
170
+ 0.5,
171
+ 0.5,
172
+ 0.5,
173
+ 0.5,
174
+ 0.5,
175
+ 0.5,
176
+ 0.5,
177
+ 0.5,
178
+ 0.5,
179
+ 0.5,
180
+ 0.5,
181
+ 0.5,
182
+ 0.5,
183
+ 0.5,
184
+ 0.5,
185
+ 0.5,
186
+ 0.5,
187
+ 0.5
188
+ ],
189
+ "build_args": {
190
+ "in_dim": 5120,
191
+ "proj_dim": 768,
192
+ "n_layers": 2,
193
+ "n_heads": 8,
194
+ "ffn_mult": 2,
195
+ "dropout": 0.1,
196
+ "num_queries": 4,
197
+ "num_classes": 20,
198
+ "feat_dim": 0,
199
+ "head_dropout": 0.1
200
+ },
201
+ "hf_prefix": "spec2mol.s1_omni_raman_linear.",
202
+ "tensor_count": 46
203
+ },
204
+ "s1_omni_uv_linear": {
205
+ "prefix": "s1_omni_uv_linear.",
206
+ "modality": "uv",
207
+ "num_classes": 20,
208
+ "vocab": [
209
+ "alkane",
210
+ "alkene",
211
+ "alkyne",
212
+ "amine",
213
+ "imine",
214
+ "nitrile",
215
+ "alcohol",
216
+ "ether",
217
+ "haloalkane",
218
+ "aldehyde",
219
+ "ketone",
220
+ "ester",
221
+ "amide",
222
+ "arene",
223
+ "imidazole",
224
+ "pyrazole",
225
+ "oxazole",
226
+ "isoxazole",
227
+ "cyclopropane",
228
+ "epoxide"
229
+ ],
230
+ "threshold_source": "0.5",
231
+ "thresholds": [
232
+ 0.5,
233
+ 0.5,
234
+ 0.5,
235
+ 0.5,
236
+ 0.5,
237
+ 0.5,
238
+ 0.5,
239
+ 0.5,
240
+ 0.5,
241
+ 0.5,
242
+ 0.5,
243
+ 0.5,
244
+ 0.5,
245
+ 0.5,
246
+ 0.5,
247
+ 0.5,
248
+ 0.5,
249
+ 0.5,
250
+ 0.5,
251
+ 0.5
252
+ ],
253
+ "build_args": {
254
+ "in_dim": 5120,
255
+ "proj_dim": 768,
256
+ "n_layers": 2,
257
+ "n_heads": 8,
258
+ "ffn_mult": 2,
259
+ "dropout": 0.1,
260
+ "num_queries": 4,
261
+ "num_classes": 20,
262
+ "feat_dim": 0,
263
+ "head_dropout": 0.1
264
+ },
265
+ "hf_prefix": "spec2mol.s1_omni_uv_linear.",
266
+ "tensor_count": 46
267
+ },
268
+ "atomcount": {
269
+ "prefix": "atomcount.",
270
+ "build_args": {
271
+ "in_dim": 256,
272
+ "hidden": [
273
+ 512,
274
+ 256
275
+ ],
276
+ "dropout": 0.1,
277
+ "out_dim": 27
278
+ },
279
+ "hf_prefix": "spec2mol.atomcount.",
280
+ "tensor_count": 8
281
+ },
282
+ "motif": {
283
+ "prefix": "motif.",
284
+ "build_args": {
285
+ "in_dim": 256,
286
+ "num_classes": 20,
287
+ "hidden": [
288
+ 512,
289
+ 256
290
+ ],
291
+ "dropout": 0.1,
292
+ "vocab": [
293
+ "alkane",
294
+ "alkene",
295
+ "alkyne",
296
+ "amine",
297
+ "imine",
298
+ "nitrile",
299
+ "alcohol",
300
+ "ether",
301
+ "haloalkane",
302
+ "aldehyde",
303
+ "ketone",
304
+ "ester",
305
+ "amide",
306
+ "arene",
307
+ "imidazole",
308
+ "pyrazole",
309
+ "oxazole",
310
+ "isoxazole",
311
+ "cyclopropane",
312
+ "epoxide"
313
+ ],
314
+ "thresholds_default": [
315
+ 0.5,
316
+ 0.5,
317
+ 0.5,
318
+ 0.5,
319
+ 0.5,
320
+ 0.5,
321
+ 0.5,
322
+ 0.5,
323
+ 0.5,
324
+ 0.5,
325
+ 0.5,
326
+ 0.5,
327
+ 0.5,
328
+ 0.5,
329
+ 0.5,
330
+ 0.5,
331
+ 0.5,
332
+ 0.5,
333
+ 0.5,
334
+ 0.5
335
+ ],
336
+ "thresholds_tuned": [
337
+ 0.09999999403953552,
338
+ 0.3499999940395355,
339
+ 0.14999999105930328,
340
+ 0.6499999761581421,
341
+ 0.8999999761581421,
342
+ 0.44999998807907104,
343
+ 0.3499999940395355,
344
+ 0.6499999761581421,
345
+ 0.699999988079071,
346
+ 0.29999998211860657,
347
+ 0.5,
348
+ 0.949999988079071,
349
+ 0.949999988079071,
350
+ 0.949999988079071,
351
+ 0.949999988079071,
352
+ 0.699999988079071,
353
+ 0.8500000238418579,
354
+ 0.800000011920929,
355
+ 0.75,
356
+ 0.699999988079071
357
+ ],
358
+ "thresholds": [
359
+ 0.09999999403953552,
360
+ 0.3499999940395355,
361
+ 0.14999999105930328,
362
+ 0.6499999761581421,
363
+ 0.8999999761581421,
364
+ 0.44999998807907104,
365
+ 0.3499999940395355,
366
+ 0.6499999761581421,
367
+ 0.699999988079071,
368
+ 0.29999998211860657,
369
+ 0.5,
370
+ 0.949999988079071,
371
+ 0.949999988079071,
372
+ 0.949999988079071,
373
+ 0.949999988079071,
374
+ 0.699999988079071,
375
+ 0.8500000238418579,
376
+ 0.800000011920929,
377
+ 0.75,
378
+ 0.699999988079071
379
+ ],
380
+ "threshold_source": "thresholds_tuned"
381
+ },
382
+ "vocab": [
383
+ "alkane",
384
+ "alkene",
385
+ "alkyne",
386
+ "amine",
387
+ "imine",
388
+ "nitrile",
389
+ "alcohol",
390
+ "ether",
391
+ "haloalkane",
392
+ "aldehyde",
393
+ "ketone",
394
+ "ester",
395
+ "amide",
396
+ "arene",
397
+ "imidazole",
398
+ "pyrazole",
399
+ "oxazole",
400
+ "isoxazole",
401
+ "cyclopropane",
402
+ "epoxide"
403
+ ],
404
+ "thresholds": [
405
+ 0.09999999403953552,
406
+ 0.3499999940395355,
407
+ 0.14999999105930328,
408
+ 0.6499999761581421,
409
+ 0.8999999761581421,
410
+ 0.44999998807907104,
411
+ 0.3499999940395355,
412
+ 0.6499999761581421,
413
+ 0.699999988079071,
414
+ 0.29999998211860657,
415
+ 0.5,
416
+ 0.949999988079071,
417
+ 0.949999988079071,
418
+ 0.949999988079071,
419
+ 0.949999988079071,
420
+ 0.699999988079071,
421
+ 0.8500000238418579,
422
+ 0.800000011920929,
423
+ 0.75,
424
+ 0.699999988079071
425
+ ],
426
+ "threshold_source": "thresholds_tuned",
427
+ "hf_prefix": "spec2mol.motif.",
428
+ "tensor_count": 8
429
+ },
430
+ "dmt": {
431
+ "prefix": "dmt.",
432
+ "config_sha256": "058051516ea7471868a028278121f4e213769ddc2835a9a0e2d5c68ddbd89399",
433
+ "keep_module_prefix": true,
434
+ "hf_prefix": "spec2mol.dmt.",
435
+ "tensor_count": 443
436
+ }
437
+ },
438
+ "source": "model.safetensors.index.json",
439
+ "shard": "model-00015-of-00021.safetensors"
440
+ },
441
+ "runtime": {
442
+ "module_source": "hf_safetensors",
443
+ "index_file": "model.safetensors.index.json",
444
+ "tensor_prefix": "spec2mol.",
445
+ "predictors": {
446
+ "ir": "hf_safetensors:spec2mol.s1_omni_ir_linear.",
447
+ "raman": "hf_safetensors:spec2mol.s1_omni_raman_linear.",
448
+ "uv": "hf_safetensors:spec2mol.s1_omni_uv_linear."
449
+ },
450
+ "route_image_order": [
451
+ "ir",
452
+ "raman",
453
+ "uv"
454
+ ],
455
+ "modality_prompts": {
456
+ "ir": "请你分析这张IR谱图",
457
+ "raman": "请你分析这张Raman谱图",
458
+ "uv": "请你分析这张UV谱图"
459
+ },
460
+ "missing_modality_policy": {
461
+ "predictor": "skip_missing",
462
+ "spcformer_context": "zero_fill_missing"
463
+ },
464
+ "motif_fusion": "onehot_intersection",
465
+ "spcformer_context_order": [
466
+ "uv",
467
+ "ir",
468
+ "raman"
469
+ ],
470
+ "motif_source": "s1_omni_linear_present_modalities",
471
+ "specformer_usage": [
472
+ "dmt_context",
473
+ "feat256",
474
+ "atom_count"
475
+ ],
476
+ "specformer_motif_head_used": false
477
+ }
478
+ }
479
+ }
configuration.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"framework":"Pytorch","task":"image-text-to-text"}
esm2/config.json ADDED
@@ -0,0 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "transformers_version": "5.6.2",
3
+ "architectures": [
4
+ "EsmForMaskedLM"
5
+ ],
6
+ "output_hidden_states": false,
7
+ "return_dict": true,
8
+ "dtype": "bfloat16",
9
+ "chunk_size_feed_forward": 0,
10
+ "is_encoder_decoder": false,
11
+ "id2label": {
12
+ "0": "LABEL_0",
13
+ "1": "LABEL_1"
14
+ },
15
+ "label2id": {
16
+ "LABEL_0": 0,
17
+ "LABEL_1": 1
18
+ },
19
+ "problem_type": null,
20
+ "vocab_size": 33,
21
+ "mask_token_id": 32,
22
+ "pad_token_id": 1,
23
+ "hidden_size": 2560,
24
+ "num_hidden_layers": 36,
25
+ "num_attention_heads": 40,
26
+ "intermediate_size": 10240,
27
+ "hidden_dropout_prob": 0.0,
28
+ "attention_probs_dropout_prob": 0.0,
29
+ "max_position_embeddings": 1026,
30
+ "rope_theta": 10000.0,
31
+ "initializer_range": 0.02,
32
+ "layer_norm_eps": 1e-05,
33
+ "position_embedding_type": "rotary",
34
+ "use_cache": true,
35
+ "emb_layer_norm_before": false,
36
+ "token_dropout": true,
37
+ "is_folding_model": false,
38
+ "esmfold_config": null,
39
+ "vocab_list": null,
40
+ "is_decoder": false,
41
+ "add_cross_attention": false,
42
+ "tie_word_embeddings": true,
43
+ "bos_token_id": null,
44
+ "eos_token_id": 2,
45
+ "classifier_dropout": null,
46
+ "hidden_act": "gelu",
47
+ "model_type": "esm",
48
+ "output_attentions": false
49
+ }
esm2/esm2_config.json ADDED
@@ -0,0 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "transformers_version": "5.6.2",
3
+ "architectures": [
4
+ "EsmForMaskedLM"
5
+ ],
6
+ "output_hidden_states": false,
7
+ "return_dict": true,
8
+ "dtype": "bfloat16",
9
+ "chunk_size_feed_forward": 0,
10
+ "is_encoder_decoder": false,
11
+ "id2label": {
12
+ "0": "LABEL_0",
13
+ "1": "LABEL_1"
14
+ },
15
+ "label2id": {
16
+ "LABEL_0": 0,
17
+ "LABEL_1": 1
18
+ },
19
+ "problem_type": null,
20
+ "vocab_size": 33,
21
+ "mask_token_id": 32,
22
+ "pad_token_id": 1,
23
+ "hidden_size": 2560,
24
+ "num_hidden_layers": 36,
25
+ "num_attention_heads": 40,
26
+ "intermediate_size": 10240,
27
+ "hidden_dropout_prob": 0.0,
28
+ "attention_probs_dropout_prob": 0.0,
29
+ "max_position_embeddings": 1026,
30
+ "rope_theta": 10000.0,
31
+ "initializer_range": 0.02,
32
+ "layer_norm_eps": 1e-05,
33
+ "position_embedding_type": "rotary",
34
+ "use_cache": true,
35
+ "emb_layer_norm_before": false,
36
+ "token_dropout": true,
37
+ "is_folding_model": false,
38
+ "esmfold_config": null,
39
+ "vocab_list": null,
40
+ "is_decoder": false,
41
+ "add_cross_attention": false,
42
+ "tie_word_embeddings": true,
43
+ "bos_token_id": null,
44
+ "eos_token_id": 2,
45
+ "classifier_dropout": null,
46
+ "hidden_act": "gelu",
47
+ "model_type": "esm",
48
+ "output_attentions": false
49
+ }
esm2/tokenizer_config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<cls>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "<eos>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "32": {
36
+ "content": "<mask>",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "backend": "custom",
45
+ "cls_token": "<cls>",
46
+ "eos_token": "<eos>",
47
+ "is_local": true,
48
+ "local_files_only": false,
49
+ "mask_token": "<mask>",
50
+ "model_max_length": 1000000000000000019884624838656,
51
+ "pad_token": "<pad>",
52
+ "tokenizer_class": "EsmTokenizer",
53
+ "unk_token": "<unk>"
54
+ }
esm2/vocab.txt ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <cls>
2
+ <pad>
3
+ <eos>
4
+ <unk>
5
+ L
6
+ A
7
+ G
8
+ V
9
+ S
10
+ E
11
+ R
12
+ T
13
+ I
14
+ D
15
+ P
16
+ K
17
+ Q
18
+ N
19
+ F
20
+ Y
21
+ M
22
+ H
23
+ W
24
+ C
25
+ X
26
+ B
27
+ U
28
+ Z
29
+ O
30
+ .
31
+ -
32
+ <null_1>
33
+ <mask>
generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "temperature": 0.8,
10
+ "top_k": 20,
11
+ "top_p": 0.95,
12
+ "transformers_version": "4.57.6"
13
+ }
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
preprocessor_config.json ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "size": {
3
+ "longest_edge": 16777216,
4
+ "shortest_edge": 65536
5
+ },
6
+ "patch_size": 16,
7
+ "temporal_patch_size": 2,
8
+ "merge_size": 2,
9
+ "image_mean": [
10
+ 0.5,
11
+ 0.5,
12
+ 0.5
13
+ ],
14
+ "image_std": [
15
+ 0.5,
16
+ 0.5,
17
+ 0.5
18
+ ],
19
+ "processor_class": "Qwen3VLProcessor",
20
+ "image_processor_type": "Qwen2VLImageProcessorFast"
21
+ }
processor_config.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "do_convert_rgb": true,
4
+ "do_normalize": true,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "image_mean": [
8
+ 0.5,
9
+ 0.5,
10
+ 0.5
11
+ ],
12
+ "image_processor_type": "Qwen2VLImageProcessor",
13
+ "image_std": [
14
+ 0.5,
15
+ 0.5,
16
+ 0.5
17
+ ],
18
+ "merge_size": 2,
19
+ "patch_size": 16,
20
+ "resample": 3,
21
+ "rescale_factor": 0.00392156862745098,
22
+ "size": {
23
+ "longest_edge": 16777216,
24
+ "shortest_edge": 65536
25
+ },
26
+ "temporal_patch_size": 2
27
+ },
28
+ "processor_class": "Qwen3VLProcessor",
29
+ "video_processor": {
30
+ "data_format": "channels_first",
31
+ "default_to_square": true,
32
+ "do_convert_rgb": true,
33
+ "do_normalize": true,
34
+ "do_rescale": true,
35
+ "do_resize": true,
36
+ "do_sample_frames": true,
37
+ "fps": 2,
38
+ "image_mean": [
39
+ 0.5,
40
+ 0.5,
41
+ 0.5
42
+ ],
43
+ "image_std": [
44
+ 0.5,
45
+ 0.5,
46
+ 0.5
47
+ ],
48
+ "max_frames": 768,
49
+ "merge_size": 2,
50
+ "min_frames": 4,
51
+ "patch_size": 16,
52
+ "resample": 3,
53
+ "rescale_factor": 0.00392156862745098,
54
+ "return_metadata": false,
55
+ "size": {
56
+ "longest_edge": 25165824,
57
+ "shortest_edge": 4096
58
+ },
59
+ "temporal_patch_size": 2,
60
+ "video_processor_type": "Qwen3VLVideoProcessor"
61
+ }
62
+ }
s1_omni_image_config.json ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "s1-omni-image-branch",
3
+ "format_version": 1,
4
+ "architectures": [
5
+ "S1OmniImageBranch"
6
+ ],
7
+ "prefixes": {
8
+ "alignment_mlp": "alignment_mlp.",
9
+ "dit": "dit.",
10
+ "vae": "vae."
11
+ },
12
+ "qwen3_vl_config": {
13
+ "architectures": [
14
+ "Qwen3VLForConditionalGeneration"
15
+ ],
16
+ "dtype": "bfloat16",
17
+ "eos_token_id": 151645,
18
+ "hidden_size": 5120,
19
+ "image_token_id": 151655,
20
+ "model_type": "qwen3_vl",
21
+ "pad_token_id": 151643,
22
+ "text_config": {
23
+ "attention_bias": false,
24
+ "attention_dropout": 0.0,
25
+ "bos_token_id": 151643,
26
+ "dtype": "bfloat16",
27
+ "eos_token_id": 151645,
28
+ "head_dim": 128,
29
+ "hidden_act": "silu",
30
+ "hidden_size": 5120,
31
+ "initializer_range": 0.02,
32
+ "intermediate_size": 25600,
33
+ "max_position_embeddings": 262144,
34
+ "model_type": "qwen3_vl_text",
35
+ "num_attention_heads": 64,
36
+ "num_hidden_layers": 64,
37
+ "num_key_value_heads": 8,
38
+ "pad_token_id": 151643,
39
+ "rms_norm_eps": 1e-06,
40
+ "rope_scaling": {
41
+ "mrope_interleaved": true,
42
+ "mrope_section": [
43
+ 24,
44
+ 20,
45
+ 20
46
+ ],
47
+ "rope_type": "default"
48
+ },
49
+ "rope_theta": 5000000,
50
+ "use_cache": false,
51
+ "vocab_size": 151936
52
+ },
53
+ "tie_word_embeddings": false,
54
+ "transformers_version": "4.57.6",
55
+ "video_token_id": 151656,
56
+ "vision_config": {
57
+ "deepstack_visual_indexes": [
58
+ 8,
59
+ 16,
60
+ 24
61
+ ],
62
+ "depth": 27,
63
+ "dtype": "bfloat16",
64
+ "hidden_act": "gelu_pytorch_tanh",
65
+ "hidden_size": 1152,
66
+ "in_channels": 3,
67
+ "initializer_range": 0.02,
68
+ "intermediate_size": 4304,
69
+ "model_type": "qwen3_vl",
70
+ "num_heads": 16,
71
+ "num_position_embeddings": 2304,
72
+ "out_hidden_size": 5120,
73
+ "pad_token_id": 151643,
74
+ "patch_size": 16,
75
+ "spatial_merge_size": 2,
76
+ "temporal_patch_size": 2
77
+ },
78
+ "vision_end_token_id": 151653,
79
+ "vision_start_token_id": 151652
80
+ },
81
+ "transformer_config": {
82
+ "_class_name": "QwenImageTransformer2DModel",
83
+ "_diffusers_version": "0.36.0.dev0",
84
+ "attention_head_dim": 128,
85
+ "axes_dims_rope": [
86
+ 16,
87
+ 56,
88
+ 56
89
+ ],
90
+ "guidance_embeds": false,
91
+ "in_channels": 64,
92
+ "joint_attention_dim": 3584,
93
+ "num_attention_heads": 24,
94
+ "num_layers": 60,
95
+ "out_channels": 16,
96
+ "patch_size": 2,
97
+ "zero_cond_t": true
98
+ },
99
+ "vae_config": {
100
+ "_class_name": "AutoencoderKLQwenImage",
101
+ "_diffusers_version": "0.36.0.dev0",
102
+ "attn_scales": [],
103
+ "base_dim": 96,
104
+ "dim_mult": [
105
+ 1,
106
+ 2,
107
+ 4,
108
+ 4
109
+ ],
110
+ "dropout": 0.0,
111
+ "latents_mean": [
112
+ -0.7571,
113
+ -0.7089,
114
+ -0.9113,
115
+ 0.1075,
116
+ -0.1745,
117
+ 0.9653,
118
+ -0.1517,
119
+ 1.5508,
120
+ 0.4134,
121
+ -0.0715,
122
+ 0.5517,
123
+ -0.3632,
124
+ -0.1922,
125
+ -0.9497,
126
+ 0.2503,
127
+ -0.2921
128
+ ],
129
+ "latents_std": [
130
+ 2.8184,
131
+ 1.4541,
132
+ 2.3275,
133
+ 2.6558,
134
+ 1.2196,
135
+ 1.7708,
136
+ 2.6052,
137
+ 2.0743,
138
+ 3.2687,
139
+ 2.1526,
140
+ 2.8652,
141
+ 1.5579,
142
+ 1.6382,
143
+ 1.1253,
144
+ 2.8251,
145
+ 1.916
146
+ ],
147
+ "num_res_blocks": 2,
148
+ "temperal_downsample": [
149
+ false,
150
+ true,
151
+ true
152
+ ],
153
+ "z_dim": 16
154
+ },
155
+ "scheduler_config": {
156
+ "_class_name": "FlowMatchEulerDiscreteScheduler",
157
+ "_diffusers_version": "0.36.0.dev0",
158
+ "base_image_seq_len": 256,
159
+ "base_shift": 0.5,
160
+ "invert_sigmas": false,
161
+ "max_image_seq_len": 8192,
162
+ "max_shift": 0.9,
163
+ "num_train_timesteps": 1000,
164
+ "shift": 1.0,
165
+ "shift_terminal": 0.02,
166
+ "stochastic_sampling": false,
167
+ "time_shift_type": "exponential",
168
+ "use_beta_sigmas": false,
169
+ "use_dynamic_shifting": true,
170
+ "use_exponential_sigmas": false,
171
+ "use_karras_sigmas": false
172
+ },
173
+ "alignment_mlp": {
174
+ "type": "mlp",
175
+ "input_dim": 5120,
176
+ "hidden_dim": 5120,
177
+ "output_dim": 3584
178
+ },
179
+ "special_tokens": {
180
+ "think_start": "<think>",
181
+ "think_end": "</think>",
182
+ "image_gen_start": "<image_gen>",
183
+ "image_gen_end": "</image_gen>",
184
+ "image_edit_start": "<image_edit>",
185
+ "image_edit_end": "</image_edit>"
186
+ },
187
+ "generation": {
188
+ "max_new_tokens": 2048,
189
+ "temperature": 0.7,
190
+ "top_p": 0.9,
191
+ "do_sample": true
192
+ },
193
+ "image_generation": {
194
+ "default_height": 1024,
195
+ "default_width": 1024,
196
+ "num_inference_steps": 50,
197
+ "guidance_scale": 1.0,
198
+ "true_cfg_scale": 4.0
199
+ },
200
+ "system_prompt": "你是中国科学院磐石团队训练的 S1-Omni 统一架构多模态模型，具备多模态图像理解和生成能力。",
201
+ "validation": {
202
+ "dit_tensors": 1933,
203
+ "alignment_mlp_tensors": 4,
204
+ "dit_missing_from_joint": 0,
205
+ "dit_extra_in_joint": 0,
206
+ "dit_shape_mismatch": 0
207
+ }
208
+ }
s1_protein_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "use_esm2": true,
3
+ "esm_fusion_dim": 512,
4
+ "esm_num_attention_heads": 8,
5
+ "esm_fusion_num_layers": 8,
6
+ "esm_fusion_ffn_dim": 2048,
7
+ "esm_unfreeze_last_n_layers": 8,
8
+ "esm_unfreeze_pooler": false,
9
+ "esm_unfreeze_final_layer_norm": true,
10
+ "esm_lr_multiplier": 0.1,
11
+ "positive_loss_weight": 20.0,
12
+ "protein_loss_type": "bce",
13
+ "asl_gamma_pos": 0.0,
14
+ "asl_gamma_neg": 4.0,
15
+ "asl_clip": 0.05,
16
+ "asl_eps": 1e-08
17
+ }
simplefold.safetensors.index.json ADDED
@@ -0,0 +1,418 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "simplefold_config": "simplefold_config.json",
4
+ "simplefold_prefix": "simplefold.",
5
+ "total_size": 2816544676
6
+ },
7
+ "weight_map": {
8
+ "simplefold.atom2latent_proj.0.bias": "model-00021-of-00021.safetensors",
9
+ "simplefold.atom2latent_proj.0.weight": "model-00021-of-00021.safetensors",
10
+ "simplefold.atom2latent_proj.1.bias": "model-00021-of-00021.safetensors",
11
+ "simplefold.atom2latent_proj.1.weight": "model-00021-of-00021.safetensors",
12
+ "simplefold.atom_dec_cond_proj.0.bias": "model-00021-of-00021.safetensors",
13
+ "simplefold.atom_dec_cond_proj.0.weight": "model-00021-of-00021.safetensors",
14
+ "simplefold.atom_dec_cond_proj.1.bias": "model-00021-of-00021.safetensors",
15
+ "simplefold.atom_dec_cond_proj.1.weight": "model-00021-of-00021.safetensors",
16
+ "simplefold.atom_decoder_transformer.blocks.0.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
17
+ "simplefold.atom_decoder_transformer.blocks.0.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
18
+ "simplefold.atom_decoder_transformer.blocks.0.attn.k_norm.scale": "model-00021-of-00021.safetensors",
19
+ "simplefold.atom_decoder_transformer.blocks.0.attn.proj.bias": "model-00021-of-00021.safetensors",
20
+ "simplefold.atom_decoder_transformer.blocks.0.attn.proj.weight": "model-00021-of-00021.safetensors",
21
+ "simplefold.atom_decoder_transformer.blocks.0.attn.q_norm.scale": "model-00021-of-00021.safetensors",
22
+ "simplefold.atom_decoder_transformer.blocks.0.attn.qkv.weight": "model-00021-of-00021.safetensors",
23
+ "simplefold.atom_decoder_transformer.blocks.0.mlp.w1.weight": "model-00021-of-00021.safetensors",
24
+ "simplefold.atom_decoder_transformer.blocks.0.mlp.w2.bias": "model-00021-of-00021.safetensors",
25
+ "simplefold.atom_decoder_transformer.blocks.0.mlp.w2.weight": "model-00021-of-00021.safetensors",
26
+ "simplefold.atom_decoder_transformer.blocks.0.mlp.w3.weight": "model-00021-of-00021.safetensors",
27
+ "simplefold.atom_decoder_transformer.blocks.1.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
28
+ "simplefold.atom_decoder_transformer.blocks.1.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
29
+ "simplefold.atom_decoder_transformer.blocks.1.attn.k_norm.scale": "model-00021-of-00021.safetensors",
30
+ "simplefold.atom_decoder_transformer.blocks.1.attn.proj.bias": "model-00021-of-00021.safetensors",
31
+ "simplefold.atom_decoder_transformer.blocks.1.attn.proj.weight": "model-00021-of-00021.safetensors",
32
+ "simplefold.atom_decoder_transformer.blocks.1.attn.q_norm.scale": "model-00021-of-00021.safetensors",
33
+ "simplefold.atom_decoder_transformer.blocks.1.attn.qkv.weight": "model-00021-of-00021.safetensors",
34
+ "simplefold.atom_decoder_transformer.blocks.1.mlp.w1.weight": "model-00021-of-00021.safetensors",
35
+ "simplefold.atom_decoder_transformer.blocks.1.mlp.w2.bias": "model-00021-of-00021.safetensors",
36
+ "simplefold.atom_decoder_transformer.blocks.1.mlp.w2.weight": "model-00021-of-00021.safetensors",
37
+ "simplefold.atom_decoder_transformer.blocks.1.mlp.w3.weight": "model-00021-of-00021.safetensors",
38
+ "simplefold.atom_enc_cond_proj.0.bias": "model-00021-of-00021.safetensors",
39
+ "simplefold.atom_enc_cond_proj.0.weight": "model-00021-of-00021.safetensors",
40
+ "simplefold.atom_enc_cond_proj.1.bias": "model-00021-of-00021.safetensors",
41
+ "simplefold.atom_enc_cond_proj.1.weight": "model-00021-of-00021.safetensors",
42
+ "simplefold.atom_encoder_transformer.blocks.0.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
43
+ "simplefold.atom_encoder_transformer.blocks.0.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
44
+ "simplefold.atom_encoder_transformer.blocks.0.attn.k_norm.scale": "model-00021-of-00021.safetensors",
45
+ "simplefold.atom_encoder_transformer.blocks.0.attn.proj.bias": "model-00021-of-00021.safetensors",
46
+ "simplefold.atom_encoder_transformer.blocks.0.attn.proj.weight": "model-00021-of-00021.safetensors",
47
+ "simplefold.atom_encoder_transformer.blocks.0.attn.q_norm.scale": "model-00021-of-00021.safetensors",
48
+ "simplefold.atom_encoder_transformer.blocks.0.attn.qkv.weight": "model-00021-of-00021.safetensors",
49
+ "simplefold.atom_encoder_transformer.blocks.0.mlp.w1.weight": "model-00021-of-00021.safetensors",
50
+ "simplefold.atom_encoder_transformer.blocks.0.mlp.w2.bias": "model-00021-of-00021.safetensors",
51
+ "simplefold.atom_encoder_transformer.blocks.0.mlp.w2.weight": "model-00021-of-00021.safetensors",
52
+ "simplefold.atom_encoder_transformer.blocks.0.mlp.w3.weight": "model-00021-of-00021.safetensors",
53
+ "simplefold.atom_encoder_transformer.blocks.1.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
54
+ "simplefold.atom_encoder_transformer.blocks.1.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
55
+ "simplefold.atom_encoder_transformer.blocks.1.attn.k_norm.scale": "model-00021-of-00021.safetensors",
56
+ "simplefold.atom_encoder_transformer.blocks.1.attn.proj.bias": "model-00021-of-00021.safetensors",
57
+ "simplefold.atom_encoder_transformer.blocks.1.attn.proj.weight": "model-00021-of-00021.safetensors",
58
+ "simplefold.atom_encoder_transformer.blocks.1.attn.q_norm.scale": "model-00021-of-00021.safetensors",
59
+ "simplefold.atom_encoder_transformer.blocks.1.attn.qkv.weight": "model-00021-of-00021.safetensors",
60
+ "simplefold.atom_encoder_transformer.blocks.1.mlp.w1.weight": "model-00021-of-00021.safetensors",
61
+ "simplefold.atom_encoder_transformer.blocks.1.mlp.w2.bias": "model-00021-of-00021.safetensors",
62
+ "simplefold.atom_encoder_transformer.blocks.1.mlp.w2.weight": "model-00021-of-00021.safetensors",
63
+ "simplefold.atom_encoder_transformer.blocks.1.mlp.w3.weight": "model-00021-of-00021.safetensors",
64
+ "simplefold.atom_feat_proj.0.bias": "model-00021-of-00021.safetensors",
65
+ "simplefold.atom_feat_proj.0.weight": "model-00021-of-00021.safetensors",
66
+ "simplefold.atom_feat_proj.1.bias": "model-00021-of-00021.safetensors",
67
+ "simplefold.atom_feat_proj.1.weight": "model-00021-of-00021.safetensors",
68
+ "simplefold.atom_in_proj.weight": "model-00021-of-00021.safetensors",
69
+ "simplefold.atom_pos_proj.weight": "model-00021-of-00021.safetensors",
70
+ "simplefold.context2atom_proj.0.bias": "model-00021-of-00021.safetensors",
71
+ "simplefold.context2atom_proj.0.weight": "model-00021-of-00021.safetensors",
72
+ "simplefold.context2atom_proj.1.bias": "model-00021-of-00021.safetensors",
73
+ "simplefold.context2atom_proj.1.weight": "model-00021-of-00021.safetensors",
74
+ "simplefold.esm_cat_proj.bias": "model-00021-of-00021.safetensors",
75
+ "simplefold.esm_cat_proj.weight": "model-00021-of-00021.safetensors",
76
+ "simplefold.esm_s_combine": "model-00021-of-00021.safetensors",
77
+ "simplefold.esm_s_proj.null_token": "model-00021-of-00021.safetensors",
78
+ "simplefold.esm_s_proj.proj.0.bias": "model-00021-of-00021.safetensors",
79
+ "simplefold.esm_s_proj.proj.0.weight": "model-00021-of-00021.safetensors",
80
+ "simplefold.esm_s_proj.proj.1.bias": "model-00021-of-00021.safetensors",
81
+ "simplefold.esm_s_proj.proj.1.weight": "model-00021-of-00021.safetensors",
82
+ "simplefold.final_layer.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
83
+ "simplefold.final_layer.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
84
+ "simplefold.final_layer.linear.bias": "model-00021-of-00021.safetensors",
85
+ "simplefold.final_layer.linear.weight": "model-00021-of-00021.safetensors",
86
+ "simplefold.latent2atom_proj.0.bias": "model-00021-of-00021.safetensors",
87
+ "simplefold.latent2atom_proj.0.weight": "model-00021-of-00021.safetensors",
88
+ "simplefold.latent2atom_proj.2.bias": "model-00021-of-00021.safetensors",
89
+ "simplefold.latent2atom_proj.2.weight": "model-00021-of-00021.safetensors",
90
+ "simplefold.latent2atom_proj.3.bias": "model-00021-of-00021.safetensors",
91
+ "simplefold.latent2atom_proj.3.weight": "model-00021-of-00021.safetensors",
92
+ "simplefold.length_embedder.0.weight": "model-00021-of-00021.safetensors",
93
+ "simplefold.length_embedder.1.bias": "model-00021-of-00021.safetensors",
94
+ "simplefold.length_embedder.1.weight": "model-00021-of-00021.safetensors",
95
+ "simplefold.llm2time_linear.bias": "model-00021-of-00021.safetensors",
96
+ "simplefold.llm2time_linear.weight": "model-00021-of-00021.safetensors",
97
+ "simplefold.pos_embedder.freq_bands": "model-00021-of-00021.safetensors",
98
+ "simplefold.text_feat_proj.0.bias": "model-00021-of-00021.safetensors",
99
+ "simplefold.text_feat_proj.0.weight": "model-00021-of-00021.safetensors",
100
+ "simplefold.text_feat_proj.2.bias": "model-00021-of-00021.safetensors",
101
+ "simplefold.text_feat_proj.2.weight": "model-00021-of-00021.safetensors",
102
+ "simplefold.text_feat_proj.3.bias": "model-00021-of-00021.safetensors",
103
+ "simplefold.text_feat_proj.3.weight": "model-00021-of-00021.safetensors",
104
+ "simplefold.text_gate": "model-00021-of-00021.safetensors",
105
+ "simplefold.time_embedder.mlp.0.bias": "model-00021-of-00021.safetensors",
106
+ "simplefold.time_embedder.mlp.0.weight": "model-00021-of-00021.safetensors",
107
+ "simplefold.time_embedder.mlp.2.bias": "model-00021-of-00021.safetensors",
108
+ "simplefold.time_embedder.mlp.2.weight": "model-00021-of-00021.safetensors",
109
+ "simplefold.trunk.blocks.0.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
110
+ "simplefold.trunk.blocks.0.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
111
+ "simplefold.trunk.blocks.0.attn.k_norm.scale": "model-00021-of-00021.safetensors",
112
+ "simplefold.trunk.blocks.0.attn.proj.bias": "model-00021-of-00021.safetensors",
113
+ "simplefold.trunk.blocks.0.attn.proj.weight": "model-00021-of-00021.safetensors",
114
+ "simplefold.trunk.blocks.0.attn.q_norm.scale": "model-00021-of-00021.safetensors",
115
+ "simplefold.trunk.blocks.0.attn.qkv.weight": "model-00021-of-00021.safetensors",
116
+ "simplefold.trunk.blocks.0.mlp.w1.weight": "model-00021-of-00021.safetensors",
117
+ "simplefold.trunk.blocks.0.mlp.w2.bias": "model-00021-of-00021.safetensors",
118
+ "simplefold.trunk.blocks.0.mlp.w2.weight": "model-00021-of-00021.safetensors",
119
+ "simplefold.trunk.blocks.0.mlp.w3.weight": "model-00021-of-00021.safetensors",
120
+ "simplefold.trunk.blocks.1.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
121
+ "simplefold.trunk.blocks.1.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
122
+ "simplefold.trunk.blocks.1.attn.k_norm.scale": "model-00021-of-00021.safetensors",
123
+ "simplefold.trunk.blocks.1.attn.proj.bias": "model-00021-of-00021.safetensors",
124
+ "simplefold.trunk.blocks.1.attn.proj.weight": "model-00021-of-00021.safetensors",
125
+ "simplefold.trunk.blocks.1.attn.q_norm.scale": "model-00021-of-00021.safetensors",
126
+ "simplefold.trunk.blocks.1.attn.qkv.weight": "model-00021-of-00021.safetensors",
127
+ "simplefold.trunk.blocks.1.mlp.w1.weight": "model-00021-of-00021.safetensors",
128
+ "simplefold.trunk.blocks.1.mlp.w2.bias": "model-00021-of-00021.safetensors",
129
+ "simplefold.trunk.blocks.1.mlp.w2.weight": "model-00021-of-00021.safetensors",
130
+ "simplefold.trunk.blocks.1.mlp.w3.weight": "model-00021-of-00021.safetensors",
131
+ "simplefold.trunk.blocks.10.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
132
+ "simplefold.trunk.blocks.10.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
133
+ "simplefold.trunk.blocks.10.attn.k_norm.scale": "model-00021-of-00021.safetensors",
134
+ "simplefold.trunk.blocks.10.attn.proj.bias": "model-00021-of-00021.safetensors",
135
+ "simplefold.trunk.blocks.10.attn.proj.weight": "model-00021-of-00021.safetensors",
136
+ "simplefold.trunk.blocks.10.attn.q_norm.scale": "model-00021-of-00021.safetensors",
137
+ "simplefold.trunk.blocks.10.attn.qkv.weight": "model-00021-of-00021.safetensors",
138
+ "simplefold.trunk.blocks.10.mlp.w1.weight": "model-00021-of-00021.safetensors",
139
+ "simplefold.trunk.blocks.10.mlp.w2.bias": "model-00021-of-00021.safetensors",
140
+ "simplefold.trunk.blocks.10.mlp.w2.weight": "model-00021-of-00021.safetensors",
141
+ "simplefold.trunk.blocks.10.mlp.w3.weight": "model-00021-of-00021.safetensors",
142
+ "simplefold.trunk.blocks.11.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
143
+ "simplefold.trunk.blocks.11.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
144
+ "simplefold.trunk.blocks.11.attn.k_norm.scale": "model-00021-of-00021.safetensors",
145
+ "simplefold.trunk.blocks.11.attn.proj.bias": "model-00021-of-00021.safetensors",
146
+ "simplefold.trunk.blocks.11.attn.proj.weight": "model-00021-of-00021.safetensors",
147
+ "simplefold.trunk.blocks.11.attn.q_norm.scale": "model-00021-of-00021.safetensors",
148
+ "simplefold.trunk.blocks.11.attn.qkv.weight": "model-00021-of-00021.safetensors",
149
+ "simplefold.trunk.blocks.11.mlp.w1.weight": "model-00021-of-00021.safetensors",
150
+ "simplefold.trunk.blocks.11.mlp.w2.bias": "model-00021-of-00021.safetensors",
151
+ "simplefold.trunk.blocks.11.mlp.w2.weight": "model-00021-of-00021.safetensors",
152
+ "simplefold.trunk.blocks.11.mlp.w3.weight": "model-00021-of-00021.safetensors",
153
+ "simplefold.trunk.blocks.12.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
154
+ "simplefold.trunk.blocks.12.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
155
+ "simplefold.trunk.blocks.12.attn.k_norm.scale": "model-00021-of-00021.safetensors",
156
+ "simplefold.trunk.blocks.12.attn.proj.bias": "model-00021-of-00021.safetensors",
157
+ "simplefold.trunk.blocks.12.attn.proj.weight": "model-00021-of-00021.safetensors",
158
+ "simplefold.trunk.blocks.12.attn.q_norm.scale": "model-00021-of-00021.safetensors",
159
+ "simplefold.trunk.blocks.12.attn.qkv.weight": "model-00021-of-00021.safetensors",
160
+ "simplefold.trunk.blocks.12.mlp.w1.weight": "model-00021-of-00021.safetensors",
161
+ "simplefold.trunk.blocks.12.mlp.w2.bias": "model-00021-of-00021.safetensors",
162
+ "simplefold.trunk.blocks.12.mlp.w2.weight": "model-00021-of-00021.safetensors",
163
+ "simplefold.trunk.blocks.12.mlp.w3.weight": "model-00021-of-00021.safetensors",
164
+ "simplefold.trunk.blocks.13.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
165
+ "simplefold.trunk.blocks.13.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
166
+ "simplefold.trunk.blocks.13.attn.k_norm.scale": "model-00021-of-00021.safetensors",
167
+ "simplefold.trunk.blocks.13.attn.proj.bias": "model-00021-of-00021.safetensors",
168
+ "simplefold.trunk.blocks.13.attn.proj.weight": "model-00021-of-00021.safetensors",
169
+ "simplefold.trunk.blocks.13.attn.q_norm.scale": "model-00021-of-00021.safetensors",
170
+ "simplefold.trunk.blocks.13.attn.qkv.weight": "model-00021-of-00021.safetensors",
171
+ "simplefold.trunk.blocks.13.mlp.w1.weight": "model-00021-of-00021.safetensors",
172
+ "simplefold.trunk.blocks.13.mlp.w2.bias": "model-00021-of-00021.safetensors",
173
+ "simplefold.trunk.blocks.13.mlp.w2.weight": "model-00021-of-00021.safetensors",
174
+ "simplefold.trunk.blocks.13.mlp.w3.weight": "model-00021-of-00021.safetensors",
175
+ "simplefold.trunk.blocks.14.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
176
+ "simplefold.trunk.blocks.14.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
177
+ "simplefold.trunk.blocks.14.attn.k_norm.scale": "model-00021-of-00021.safetensors",
178
+ "simplefold.trunk.blocks.14.attn.proj.bias": "model-00021-of-00021.safetensors",
179
+ "simplefold.trunk.blocks.14.attn.proj.weight": "model-00021-of-00021.safetensors",
180
+ "simplefold.trunk.blocks.14.attn.q_norm.scale": "model-00021-of-00021.safetensors",
181
+ "simplefold.trunk.blocks.14.attn.qkv.weight": "model-00021-of-00021.safetensors",
182
+ "simplefold.trunk.blocks.14.mlp.w1.weight": "model-00021-of-00021.safetensors",
183
+ "simplefold.trunk.blocks.14.mlp.w2.bias": "model-00021-of-00021.safetensors",
184
+ "simplefold.trunk.blocks.14.mlp.w2.weight": "model-00021-of-00021.safetensors",
185
+ "simplefold.trunk.blocks.14.mlp.w3.weight": "model-00021-of-00021.safetensors",
186
+ "simplefold.trunk.blocks.15.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
187
+ "simplefold.trunk.blocks.15.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
188
+ "simplefold.trunk.blocks.15.attn.k_norm.scale": "model-00021-of-00021.safetensors",
189
+ "simplefold.trunk.blocks.15.attn.proj.bias": "model-00021-of-00021.safetensors",
190
+ "simplefold.trunk.blocks.15.attn.proj.weight": "model-00021-of-00021.safetensors",
191
+ "simplefold.trunk.blocks.15.attn.q_norm.scale": "model-00021-of-00021.safetensors",
192
+ "simplefold.trunk.blocks.15.attn.qkv.weight": "model-00021-of-00021.safetensors",
193
+ "simplefold.trunk.blocks.15.mlp.w1.weight": "model-00021-of-00021.safetensors",
194
+ "simplefold.trunk.blocks.15.mlp.w2.bias": "model-00021-of-00021.safetensors",
195
+ "simplefold.trunk.blocks.15.mlp.w2.weight": "model-00021-of-00021.safetensors",
196
+ "simplefold.trunk.blocks.15.mlp.w3.weight": "model-00021-of-00021.safetensors",
197
+ "simplefold.trunk.blocks.16.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
198
+ "simplefold.trunk.blocks.16.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
199
+ "simplefold.trunk.blocks.16.attn.k_norm.scale": "model-00021-of-00021.safetensors",
200
+ "simplefold.trunk.blocks.16.attn.proj.bias": "model-00021-of-00021.safetensors",
201
+ "simplefold.trunk.blocks.16.attn.proj.weight": "model-00021-of-00021.safetensors",
202
+ "simplefold.trunk.blocks.16.attn.q_norm.scale": "model-00021-of-00021.safetensors",
203
+ "simplefold.trunk.blocks.16.attn.qkv.weight": "model-00021-of-00021.safetensors",
204
+ "simplefold.trunk.blocks.16.mlp.w1.weight": "model-00021-of-00021.safetensors",
205
+ "simplefold.trunk.blocks.16.mlp.w2.bias": "model-00021-of-00021.safetensors",
206
+ "simplefold.trunk.blocks.16.mlp.w2.weight": "model-00021-of-00021.safetensors",
207
+ "simplefold.trunk.blocks.16.mlp.w3.weight": "model-00021-of-00021.safetensors",
208
+ "simplefold.trunk.blocks.17.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
209
+ "simplefold.trunk.blocks.17.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
210
+ "simplefold.trunk.blocks.17.attn.k_norm.scale": "model-00021-of-00021.safetensors",
211
+ "simplefold.trunk.blocks.17.attn.proj.bias": "model-00021-of-00021.safetensors",
212
+ "simplefold.trunk.blocks.17.attn.proj.weight": "model-00021-of-00021.safetensors",
213
+ "simplefold.trunk.blocks.17.attn.q_norm.scale": "model-00021-of-00021.safetensors",
214
+ "simplefold.trunk.blocks.17.attn.qkv.weight": "model-00021-of-00021.safetensors",
215
+ "simplefold.trunk.blocks.17.mlp.w1.weight": "model-00021-of-00021.safetensors",
216
+ "simplefold.trunk.blocks.17.mlp.w2.bias": "model-00021-of-00021.safetensors",
217
+ "simplefold.trunk.blocks.17.mlp.w2.weight": "model-00021-of-00021.safetensors",
218
+ "simplefold.trunk.blocks.17.mlp.w3.weight": "model-00021-of-00021.safetensors",
219
+ "simplefold.trunk.blocks.18.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
220
+ "simplefold.trunk.blocks.18.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
221
+ "simplefold.trunk.blocks.18.attn.k_norm.scale": "model-00021-of-00021.safetensors",
222
+ "simplefold.trunk.blocks.18.attn.proj.bias": "model-00021-of-00021.safetensors",
223
+ "simplefold.trunk.blocks.18.attn.proj.weight": "model-00021-of-00021.safetensors",
224
+ "simplefold.trunk.blocks.18.attn.q_norm.scale": "model-00021-of-00021.safetensors",
225
+ "simplefold.trunk.blocks.18.attn.qkv.weight": "model-00021-of-00021.safetensors",
226
+ "simplefold.trunk.blocks.18.mlp.w1.weight": "model-00021-of-00021.safetensors",
227
+ "simplefold.trunk.blocks.18.mlp.w2.bias": "model-00021-of-00021.safetensors",
228
+ "simplefold.trunk.blocks.18.mlp.w2.weight": "model-00021-of-00021.safetensors",
229
+ "simplefold.trunk.blocks.18.mlp.w3.weight": "model-00021-of-00021.safetensors",
230
+ "simplefold.trunk.blocks.19.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
231
+ "simplefold.trunk.blocks.19.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
232
+ "simplefold.trunk.blocks.19.attn.k_norm.scale": "model-00021-of-00021.safetensors",
233
+ "simplefold.trunk.blocks.19.attn.proj.bias": "model-00021-of-00021.safetensors",
234
+ "simplefold.trunk.blocks.19.attn.proj.weight": "model-00021-of-00021.safetensors",
235
+ "simplefold.trunk.blocks.19.attn.q_norm.scale": "model-00021-of-00021.safetensors",
236
+ "simplefold.trunk.blocks.19.attn.qkv.weight": "model-00021-of-00021.safetensors",
237
+ "simplefold.trunk.blocks.19.mlp.w1.weight": "model-00021-of-00021.safetensors",
238
+ "simplefold.trunk.blocks.19.mlp.w2.bias": "model-00021-of-00021.safetensors",
239
+ "simplefold.trunk.blocks.19.mlp.w2.weight": "model-00021-of-00021.safetensors",
240
+ "simplefold.trunk.blocks.19.mlp.w3.weight": "model-00021-of-00021.safetensors",
241
+ "simplefold.trunk.blocks.2.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
242
+ "simplefold.trunk.blocks.2.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
243
+ "simplefold.trunk.blocks.2.attn.k_norm.scale": "model-00021-of-00021.safetensors",
244
+ "simplefold.trunk.blocks.2.attn.proj.bias": "model-00021-of-00021.safetensors",
245
+ "simplefold.trunk.blocks.2.attn.proj.weight": "model-00021-of-00021.safetensors",
246
+ "simplefold.trunk.blocks.2.attn.q_norm.scale": "model-00021-of-00021.safetensors",
247
+ "simplefold.trunk.blocks.2.attn.qkv.weight": "model-00021-of-00021.safetensors",
248
+ "simplefold.trunk.blocks.2.mlp.w1.weight": "model-00021-of-00021.safetensors",
249
+ "simplefold.trunk.blocks.2.mlp.w2.bias": "model-00021-of-00021.safetensors",
250
+ "simplefold.trunk.blocks.2.mlp.w2.weight": "model-00021-of-00021.safetensors",
251
+ "simplefold.trunk.blocks.2.mlp.w3.weight": "model-00021-of-00021.safetensors",
252
+ "simplefold.trunk.blocks.20.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
253
+ "simplefold.trunk.blocks.20.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
254
+ "simplefold.trunk.blocks.20.attn.k_norm.scale": "model-00021-of-00021.safetensors",
255
+ "simplefold.trunk.blocks.20.attn.proj.bias": "model-00021-of-00021.safetensors",
256
+ "simplefold.trunk.blocks.20.attn.proj.weight": "model-00021-of-00021.safetensors",
257
+ "simplefold.trunk.blocks.20.attn.q_norm.scale": "model-00021-of-00021.safetensors",
258
+ "simplefold.trunk.blocks.20.attn.qkv.weight": "model-00021-of-00021.safetensors",
259
+ "simplefold.trunk.blocks.20.mlp.w1.weight": "model-00021-of-00021.safetensors",
260
+ "simplefold.trunk.blocks.20.mlp.w2.bias": "model-00021-of-00021.safetensors",
261
+ "simplefold.trunk.blocks.20.mlp.w2.weight": "model-00021-of-00021.safetensors",
262
+ "simplefold.trunk.blocks.20.mlp.w3.weight": "model-00021-of-00021.safetensors",
263
+ "simplefold.trunk.blocks.21.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
264
+ "simplefold.trunk.blocks.21.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
265
+ "simplefold.trunk.blocks.21.attn.k_norm.scale": "model-00021-of-00021.safetensors",
266
+ "simplefold.trunk.blocks.21.attn.proj.bias": "model-00021-of-00021.safetensors",
267
+ "simplefold.trunk.blocks.21.attn.proj.weight": "model-00021-of-00021.safetensors",
268
+ "simplefold.trunk.blocks.21.attn.q_norm.scale": "model-00021-of-00021.safetensors",
269
+ "simplefold.trunk.blocks.21.attn.qkv.weight": "model-00021-of-00021.safetensors",
270
+ "simplefold.trunk.blocks.21.mlp.w1.weight": "model-00021-of-00021.safetensors",
271
+ "simplefold.trunk.blocks.21.mlp.w2.bias": "model-00021-of-00021.safetensors",
272
+ "simplefold.trunk.blocks.21.mlp.w2.weight": "model-00021-of-00021.safetensors",
273
+ "simplefold.trunk.blocks.21.mlp.w3.weight": "model-00021-of-00021.safetensors",
274
+ "simplefold.trunk.blocks.22.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
275
+ "simplefold.trunk.blocks.22.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
276
+ "simplefold.trunk.blocks.22.attn.k_norm.scale": "model-00021-of-00021.safetensors",
277
+ "simplefold.trunk.blocks.22.attn.proj.bias": "model-00021-of-00021.safetensors",
278
+ "simplefold.trunk.blocks.22.attn.proj.weight": "model-00021-of-00021.safetensors",
279
+ "simplefold.trunk.blocks.22.attn.q_norm.scale": "model-00021-of-00021.safetensors",
280
+ "simplefold.trunk.blocks.22.attn.qkv.weight": "model-00021-of-00021.safetensors",
281
+ "simplefold.trunk.blocks.22.mlp.w1.weight": "model-00021-of-00021.safetensors",
282
+ "simplefold.trunk.blocks.22.mlp.w2.bias": "model-00021-of-00021.safetensors",
283
+ "simplefold.trunk.blocks.22.mlp.w2.weight": "model-00021-of-00021.safetensors",
284
+ "simplefold.trunk.blocks.22.mlp.w3.weight": "model-00021-of-00021.safetensors",
285
+ "simplefold.trunk.blocks.23.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
286
+ "simplefold.trunk.blocks.23.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
287
+ "simplefold.trunk.blocks.23.attn.k_norm.scale": "model-00021-of-00021.safetensors",
288
+ "simplefold.trunk.blocks.23.attn.proj.bias": "model-00021-of-00021.safetensors",
289
+ "simplefold.trunk.blocks.23.attn.proj.weight": "model-00021-of-00021.safetensors",
290
+ "simplefold.trunk.blocks.23.attn.q_norm.scale": "model-00021-of-00021.safetensors",
291
+ "simplefold.trunk.blocks.23.attn.qkv.weight": "model-00021-of-00021.safetensors",
292
+ "simplefold.trunk.blocks.23.mlp.w1.weight": "model-00021-of-00021.safetensors",
293
+ "simplefold.trunk.blocks.23.mlp.w2.bias": "model-00021-of-00021.safetensors",
294
+ "simplefold.trunk.blocks.23.mlp.w2.weight": "model-00021-of-00021.safetensors",
295
+ "simplefold.trunk.blocks.23.mlp.w3.weight": "model-00021-of-00021.safetensors",
296
+ "simplefold.trunk.blocks.24.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
297
+ "simplefold.trunk.blocks.24.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
298
+ "simplefold.trunk.blocks.24.attn.k_norm.scale": "model-00021-of-00021.safetensors",
299
+ "simplefold.trunk.blocks.24.attn.proj.bias": "model-00021-of-00021.safetensors",
300
+ "simplefold.trunk.blocks.24.attn.proj.weight": "model-00021-of-00021.safetensors",
301
+ "simplefold.trunk.blocks.24.attn.q_norm.scale": "model-00021-of-00021.safetensors",
302
+ "simplefold.trunk.blocks.24.attn.qkv.weight": "model-00021-of-00021.safetensors",
303
+ "simplefold.trunk.blocks.24.mlp.w1.weight": "model-00021-of-00021.safetensors",
304
+ "simplefold.trunk.blocks.24.mlp.w2.bias": "model-00021-of-00021.safetensors",
305
+ "simplefold.trunk.blocks.24.mlp.w2.weight": "model-00021-of-00021.safetensors",
306
+ "simplefold.trunk.blocks.24.mlp.w3.weight": "model-00021-of-00021.safetensors",
307
+ "simplefold.trunk.blocks.25.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
308
+ "simplefold.trunk.blocks.25.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
309
+ "simplefold.trunk.blocks.25.attn.k_norm.scale": "model-00021-of-00021.safetensors",
310
+ "simplefold.trunk.blocks.25.attn.proj.bias": "model-00021-of-00021.safetensors",
311
+ "simplefold.trunk.blocks.25.attn.proj.weight": "model-00021-of-00021.safetensors",
312
+ "simplefold.trunk.blocks.25.attn.q_norm.scale": "model-00021-of-00021.safetensors",
313
+ "simplefold.trunk.blocks.25.attn.qkv.weight": "model-00021-of-00021.safetensors",
314
+ "simplefold.trunk.blocks.25.mlp.w1.weight": "model-00021-of-00021.safetensors",
315
+ "simplefold.trunk.blocks.25.mlp.w2.bias": "model-00021-of-00021.safetensors",
316
+ "simplefold.trunk.blocks.25.mlp.w2.weight": "model-00021-of-00021.safetensors",
317
+ "simplefold.trunk.blocks.25.mlp.w3.weight": "model-00021-of-00021.safetensors",
318
+ "simplefold.trunk.blocks.26.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
319
+ "simplefold.trunk.blocks.26.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
320
+ "simplefold.trunk.blocks.26.attn.k_norm.scale": "model-00021-of-00021.safetensors",
321
+ "simplefold.trunk.blocks.26.attn.proj.bias": "model-00021-of-00021.safetensors",
322
+ "simplefold.trunk.blocks.26.attn.proj.weight": "model-00021-of-00021.safetensors",
323
+ "simplefold.trunk.blocks.26.attn.q_norm.scale": "model-00021-of-00021.safetensors",
324
+ "simplefold.trunk.blocks.26.attn.qkv.weight": "model-00021-of-00021.safetensors",
325
+ "simplefold.trunk.blocks.26.mlp.w1.weight": "model-00021-of-00021.safetensors",
326
+ "simplefold.trunk.blocks.26.mlp.w2.bias": "model-00021-of-00021.safetensors",
327
+ "simplefold.trunk.blocks.26.mlp.w2.weight": "model-00021-of-00021.safetensors",
328
+ "simplefold.trunk.blocks.26.mlp.w3.weight": "model-00021-of-00021.safetensors",
329
+ "simplefold.trunk.blocks.27.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
330
+ "simplefold.trunk.blocks.27.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
331
+ "simplefold.trunk.blocks.27.attn.k_norm.scale": "model-00021-of-00021.safetensors",
332
+ "simplefold.trunk.blocks.27.attn.proj.bias": "model-00021-of-00021.safetensors",
333
+ "simplefold.trunk.blocks.27.attn.proj.weight": "model-00021-of-00021.safetensors",
334
+ "simplefold.trunk.blocks.27.attn.q_norm.scale": "model-00021-of-00021.safetensors",
335
+ "simplefold.trunk.blocks.27.attn.qkv.weight": "model-00021-of-00021.safetensors",
336
+ "simplefold.trunk.blocks.27.mlp.w1.weight": "model-00021-of-00021.safetensors",
337
+ "simplefold.trunk.blocks.27.mlp.w2.bias": "model-00021-of-00021.safetensors",
338
+ "simplefold.trunk.blocks.27.mlp.w2.weight": "model-00021-of-00021.safetensors",
339
+ "simplefold.trunk.blocks.27.mlp.w3.weight": "model-00021-of-00021.safetensors",
340
+ "simplefold.trunk.blocks.3.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
341
+ "simplefold.trunk.blocks.3.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
342
+ "simplefold.trunk.blocks.3.attn.k_norm.scale": "model-00021-of-00021.safetensors",
343
+ "simplefold.trunk.blocks.3.attn.proj.bias": "model-00021-of-00021.safetensors",
344
+ "simplefold.trunk.blocks.3.attn.proj.weight": "model-00021-of-00021.safetensors",
345
+ "simplefold.trunk.blocks.3.attn.q_norm.scale": "model-00021-of-00021.safetensors",
346
+ "simplefold.trunk.blocks.3.attn.qkv.weight": "model-00021-of-00021.safetensors",
347
+ "simplefold.trunk.blocks.3.mlp.w1.weight": "model-00021-of-00021.safetensors",
348
+ "simplefold.trunk.blocks.3.mlp.w2.bias": "model-00021-of-00021.safetensors",
349
+ "simplefold.trunk.blocks.3.mlp.w2.weight": "model-00021-of-00021.safetensors",
350
+ "simplefold.trunk.blocks.3.mlp.w3.weight": "model-00021-of-00021.safetensors",
351
+ "simplefold.trunk.blocks.4.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
352
+ "simplefold.trunk.blocks.4.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
353
+ "simplefold.trunk.blocks.4.attn.k_norm.scale": "model-00021-of-00021.safetensors",
354
+ "simplefold.trunk.blocks.4.attn.proj.bias": "model-00021-of-00021.safetensors",
355
+ "simplefold.trunk.blocks.4.attn.proj.weight": "model-00021-of-00021.safetensors",
356
+ "simplefold.trunk.blocks.4.attn.q_norm.scale": "model-00021-of-00021.safetensors",
357
+ "simplefold.trunk.blocks.4.attn.qkv.weight": "model-00021-of-00021.safetensors",
358
+ "simplefold.trunk.blocks.4.mlp.w1.weight": "model-00021-of-00021.safetensors",
359
+ "simplefold.trunk.blocks.4.mlp.w2.bias": "model-00021-of-00021.safetensors",
360
+ "simplefold.trunk.blocks.4.mlp.w2.weight": "model-00021-of-00021.safetensors",
361
+ "simplefold.trunk.blocks.4.mlp.w3.weight": "model-00021-of-00021.safetensors",
362
+ "simplefold.trunk.blocks.5.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
363
+ "simplefold.trunk.blocks.5.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
364
+ "simplefold.trunk.blocks.5.attn.k_norm.scale": "model-00021-of-00021.safetensors",
365
+ "simplefold.trunk.blocks.5.attn.proj.bias": "model-00021-of-00021.safetensors",
366
+ "simplefold.trunk.blocks.5.attn.proj.weight": "model-00021-of-00021.safetensors",
367
+ "simplefold.trunk.blocks.5.attn.q_norm.scale": "model-00021-of-00021.safetensors",
368
+ "simplefold.trunk.blocks.5.attn.qkv.weight": "model-00021-of-00021.safetensors",
369
+ "simplefold.trunk.blocks.5.mlp.w1.weight": "model-00021-of-00021.safetensors",
370
+ "simplefold.trunk.blocks.5.mlp.w2.bias": "model-00021-of-00021.safetensors",
371
+ "simplefold.trunk.blocks.5.mlp.w2.weight": "model-00021-of-00021.safetensors",
372
+ "simplefold.trunk.blocks.5.mlp.w3.weight": "model-00021-of-00021.safetensors",
373
+ "simplefold.trunk.blocks.6.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
374
+ "simplefold.trunk.blocks.6.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
375
+ "simplefold.trunk.blocks.6.attn.k_norm.scale": "model-00021-of-00021.safetensors",
376
+ "simplefold.trunk.blocks.6.attn.proj.bias": "model-00021-of-00021.safetensors",
377
+ "simplefold.trunk.blocks.6.attn.proj.weight": "model-00021-of-00021.safetensors",
378
+ "simplefold.trunk.blocks.6.attn.q_norm.scale": "model-00021-of-00021.safetensors",
379
+ "simplefold.trunk.blocks.6.attn.qkv.weight": "model-00021-of-00021.safetensors",
380
+ "simplefold.trunk.blocks.6.mlp.w1.weight": "model-00021-of-00021.safetensors",
381
+ "simplefold.trunk.blocks.6.mlp.w2.bias": "model-00021-of-00021.safetensors",
382
+ "simplefold.trunk.blocks.6.mlp.w2.weight": "model-00021-of-00021.safetensors",
383
+ "simplefold.trunk.blocks.6.mlp.w3.weight": "model-00021-of-00021.safetensors",
384
+ "simplefold.trunk.blocks.7.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
385
+ "simplefold.trunk.blocks.7.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
386
+ "simplefold.trunk.blocks.7.attn.k_norm.scale": "model-00021-of-00021.safetensors",
387
+ "simplefold.trunk.blocks.7.attn.proj.bias": "model-00021-of-00021.safetensors",
388
+ "simplefold.trunk.blocks.7.attn.proj.weight": "model-00021-of-00021.safetensors",
389
+ "simplefold.trunk.blocks.7.attn.q_norm.scale": "model-00021-of-00021.safetensors",
390
+ "simplefold.trunk.blocks.7.attn.qkv.weight": "model-00021-of-00021.safetensors",
391
+ "simplefold.trunk.blocks.7.mlp.w1.weight": "model-00021-of-00021.safetensors",
392
+ "simplefold.trunk.blocks.7.mlp.w2.bias": "model-00021-of-00021.safetensors",
393
+ "simplefold.trunk.blocks.7.mlp.w2.weight": "model-00021-of-00021.safetensors",
394
+ "simplefold.trunk.blocks.7.mlp.w3.weight": "model-00021-of-00021.safetensors",
395
+ "simplefold.trunk.blocks.8.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
396
+ "simplefold.trunk.blocks.8.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
397
+ "simplefold.trunk.blocks.8.attn.k_norm.scale": "model-00021-of-00021.safetensors",
398
+ "simplefold.trunk.blocks.8.attn.proj.bias": "model-00021-of-00021.safetensors",
399
+ "simplefold.trunk.blocks.8.attn.proj.weight": "model-00021-of-00021.safetensors",
400
+ "simplefold.trunk.blocks.8.attn.q_norm.scale": "model-00021-of-00021.safetensors",
401
+ "simplefold.trunk.blocks.8.attn.qkv.weight": "model-00021-of-00021.safetensors",
402
+ "simplefold.trunk.blocks.8.mlp.w1.weight": "model-00021-of-00021.safetensors",
403
+ "simplefold.trunk.blocks.8.mlp.w2.bias": "model-00021-of-00021.safetensors",
404
+ "simplefold.trunk.blocks.8.mlp.w2.weight": "model-00021-of-00021.safetensors",
405
+ "simplefold.trunk.blocks.8.mlp.w3.weight": "model-00021-of-00021.safetensors",
406
+ "simplefold.trunk.blocks.9.adaLN_modulation.1.bias": "model-00021-of-00021.safetensors",
407
+ "simplefold.trunk.blocks.9.adaLN_modulation.1.weight": "model-00021-of-00021.safetensors",
408
+ "simplefold.trunk.blocks.9.attn.k_norm.scale": "model-00021-of-00021.safetensors",
409
+ "simplefold.trunk.blocks.9.attn.proj.bias": "model-00021-of-00021.safetensors",
410
+ "simplefold.trunk.blocks.9.attn.proj.weight": "model-00021-of-00021.safetensors",
411
+ "simplefold.trunk.blocks.9.attn.q_norm.scale": "model-00021-of-00021.safetensors",
412
+ "simplefold.trunk.blocks.9.attn.qkv.weight": "model-00021-of-00021.safetensors",
413
+ "simplefold.trunk.blocks.9.mlp.w1.weight": "model-00021-of-00021.safetensors",
414
+ "simplefold.trunk.blocks.9.mlp.w2.bias": "model-00021-of-00021.safetensors",
415
+ "simplefold.trunk.blocks.9.mlp.w2.weight": "model-00021-of-00021.safetensors",
416
+ "simplefold.trunk.blocks.9.mlp.w3.weight": "model-00021-of-00021.safetensors"
417
+ }
418
+ }
simplefold_config.json ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_target_": "model.torch.text_embedding_architecture.FoldingDiTwithLLM",
3
+ "hidden_size": 1152,
4
+ "num_heads": 16,
5
+ "atom_num_heads": 4,
6
+ "output_channels": 3,
7
+ "use_atom_mask": false,
8
+ "use_length_condition": true,
9
+ "esm_dropout_prob": 0.0,
10
+ "esm_model": "esm2_3B",
11
+ "llm_hidden_size": 5120,
12
+ "text_proj_mode": "mlp_deep",
13
+ "text_gate_mode": "scalar",
14
+ "text_embedding_diag_interval": 0,
15
+ "time_embedder": {
16
+ "_target_": "model.torch.layers.TimestepEmbedder",
17
+ "hidden_size": 1152
18
+ },
19
+ "aminoacid_pos_embedder": {
20
+ "_target_": "model.torch.pos_embed.AbsolutePositionEncoding",
21
+ "in_dim": 1,
22
+ "embed_dim": 1152,
23
+ "include_input": true
24
+ },
25
+ "pos_embedder": {
26
+ "_target_": "model.torch.pos_embed.FourierPositionEncoding",
27
+ "in_dim": 3,
28
+ "include_input": true,
29
+ "min_freq_log2": 0,
30
+ "max_freq_log2": 12,
31
+ "num_freqs": 128,
32
+ "log_sampling": true
33
+ },
34
+ "trunk": {
35
+ "_target_": "model.torch.blocks.HomogenTrunk",
36
+ "depth": 28,
37
+ "block": {
38
+ "_target_": "model.torch.blocks.DiTBlock",
39
+ "_partial_": true,
40
+ "hidden_size": 1152,
41
+ "mlp_ratio": 4.0,
42
+ "use_swiglu": true,
43
+ "self_attention_layer": {
44
+ "_target_": "model.torch.layers.EfficientSelfAttentionLayer",
45
+ "_partial_": true,
46
+ "hidden_size": 1152,
47
+ "num_heads": 16,
48
+ "qk_norm": true,
49
+ "pos_embedder": {
50
+ "_target_": "model.torch.pos_embed.AxialRotaryPositionEncoding",
51
+ "in_dim": 4,
52
+ "embed_dim": 1152,
53
+ "num_heads": 16,
54
+ "base": 100.0
55
+ }
56
+ }
57
+ }
58
+ },
59
+ "atom_hidden_size_enc": 256,
60
+ "atom_n_queries_enc": 32,
61
+ "atom_n_keys_enc": 128,
62
+ "atom_encoder_transformer": {
63
+ "_target_": "model.torch.blocks.HomogenTrunk",
64
+ "depth": 2,
65
+ "block": {
66
+ "_target_": "model.torch.blocks.DiTBlock",
67
+ "_partial_": true,
68
+ "hidden_size": 256,
69
+ "mlp_ratio": 4.0,
70
+ "use_swiglu": true,
71
+ "self_attention_layer": {
72
+ "_target_": "model.torch.layers.EfficientSelfAttentionLayer",
73
+ "_partial_": true,
74
+ "hidden_size": 256,
75
+ "num_heads": 4,
76
+ "qk_norm": true,
77
+ "pos_embedder": {
78
+ "_target_": "model.torch.pos_embed.AxialRotaryPositionEncoding",
79
+ "in_dim": 4,
80
+ "embed_dim": 256,
81
+ "num_heads": 4,
82
+ "base": 100.0
83
+ }
84
+ }
85
+ }
86
+ },
87
+ "atom_hidden_size_dec": 256,
88
+ "atom_n_queries_dec": 32,
89
+ "atom_n_keys_dec": 128,
90
+ "atom_decoder_transformer": {
91
+ "_target_": "model.torch.blocks.HomogenTrunk",
92
+ "depth": 2,
93
+ "block": {
94
+ "_target_": "model.torch.blocks.DiTBlock",
95
+ "_partial_": true,
96
+ "hidden_size": 256,
97
+ "mlp_ratio": 4.0,
98
+ "use_swiglu": true,
99
+ "self_attention_layer": {
100
+ "_target_": "model.torch.layers.EfficientSelfAttentionLayer",
101
+ "_partial_": true,
102
+ "hidden_size": 256,
103
+ "num_heads": 4,
104
+ "qk_norm": true,
105
+ "pos_embedder": {
106
+ "_target_": "model.torch.pos_embed.AxialRotaryPositionEncoding",
107
+ "in_dim": 4,
108
+ "embed_dim": 256,
109
+ "num_heads": 4,
110
+ "base": 100.0
111
+ }
112
+ }
113
+ }
114
+ },
115
+ "architecture": "foldingdit"
116
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>",
16
+ "<image_gen>",
17
+ "<image_edit>",
18
+ "<linear_cla>",
19
+ "<linear_pre>",
20
+ "<prot_st>",
21
+ "<prot_cla>",
22
+ "<spectra_st>"
23
+ ],
24
+ "eos_token": {
25
+ "content": "<|im_end|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ },
31
+ "pad_token": {
32
+ "content": "<|endoftext|>",
33
+ "lstrip": false,
34
+ "normalized": false,
35
+ "rstrip": false,
36
+ "single_word": false
37
+ }
38
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6162f0accd42be0927413f4d56a7af551de56e729f975a5aae6665b5c29f8c0e
3
+ size 11423971
tokenizer_config.json ADDED
@@ -0,0 +1,303 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ },
181
+ "151665": {
182
+ "content": "<tool_response>",
183
+ "lstrip": false,
184
+ "normalized": false,
185
+ "rstrip": false,
186
+ "single_word": false,
187
+ "special": false
188
+ },
189
+ "151666": {
190
+ "content": "</tool_response>",
191
+ "lstrip": false,
192
+ "normalized": false,
193
+ "rstrip": false,
194
+ "single_word": false,
195
+ "special": false
196
+ },
197
+ "151667": {
198
+ "content": "<think>",
199
+ "lstrip": false,
200
+ "normalized": false,
201
+ "rstrip": false,
202
+ "single_word": false,
203
+ "special": false
204
+ },
205
+ "151668": {
206
+ "content": "</think>",
207
+ "lstrip": false,
208
+ "normalized": false,
209
+ "rstrip": false,
210
+ "single_word": false,
211
+ "special": false
212
+ },
213
+ "151669": {
214
+ "content": "<image_gen>",
215
+ "lstrip": false,
216
+ "normalized": false,
217
+ "rstrip": false,
218
+ "single_word": false,
219
+ "special": true
220
+ },
221
+ "151670": {
222
+ "content": "<image_edit>",
223
+ "lstrip": false,
224
+ "normalized": false,
225
+ "rstrip": false,
226
+ "single_word": false,
227
+ "special": true
228
+ },
229
+ "151671": {
230
+ "content": "<linear_cla>",
231
+ "lstrip": false,
232
+ "normalized": false,
233
+ "rstrip": false,
234
+ "single_word": false,
235
+ "special": true
236
+ },
237
+ "151672": {
238
+ "content": "<linear_pre>",
239
+ "lstrip": false,
240
+ "normalized": false,
241
+ "rstrip": false,
242
+ "single_word": false,
243
+ "special": true
244
+ },
245
+ "151673": {
246
+ "content": "<prot_st>",
247
+ "lstrip": false,
248
+ "normalized": false,
249
+ "rstrip": false,
250
+ "single_word": false,
251
+ "special": true
252
+ },
253
+ "151674": {
254
+ "content": "<prot_cla>",
255
+ "lstrip": false,
256
+ "normalized": false,
257
+ "rstrip": false,
258
+ "single_word": false,
259
+ "special": true
260
+ },
261
+ "151675": {
262
+ "content": "<spectra_st>",
263
+ "lstrip": false,
264
+ "normalized": false,
265
+ "rstrip": false,
266
+ "single_word": false,
267
+ "special": true
268
+ }
269
+ },
270
+ "additional_special_tokens": [
271
+ "<|im_start|>",
272
+ "<|im_end|>",
273
+ "<|object_ref_start|>",
274
+ "<|object_ref_end|>",
275
+ "<|box_start|>",
276
+ "<|box_end|>",
277
+ "<|quad_start|>",
278
+ "<|quad_end|>",
279
+ "<|vision_start|>",
280
+ "<|vision_end|>",
281
+ "<|vision_pad|>",
282
+ "<|image_pad|>",
283
+ "<|video_pad|>",
284
+ "<image_gen>",
285
+ "<image_edit>",
286
+ "<linear_cla>",
287
+ "<linear_pre>",
288
+ "<prot_st>",
289
+ "<prot_cla>",
290
+ "<spectra_st>"
291
+ ],
292
+ "bos_token": null,
293
+ "clean_up_tokenization_spaces": false,
294
+ "eos_token": "<|im_end|>",
295
+ "errors": "replace",
296
+ "extra_special_tokens": {},
297
+ "model_max_length": 262144,
298
+ "pad_token": "<|endoftext|>",
299
+ "processor_class": "Qwen3VLProcessor",
300
+ "split_special_tokens": false,
301
+ "tokenizer_class": "Qwen2Tokenizer",
302
+ "unk_token": null
303
+ }
video_preprocessor_config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": null,
3
+ "data_format": "channels_first",
4
+ "default_to_square": true,
5
+ "device": null,
6
+ "do_center_crop": null,
7
+ "do_convert_rgb": true,
8
+ "do_normalize": true,
9
+ "do_rescale": true,
10
+ "do_resize": true,
11
+ "do_sample_frames": true,
12
+ "fps": 2,
13
+ "image_mean": [
14
+ 0.5,
15
+ 0.5,
16
+ 0.5
17
+ ],
18
+ "image_std": [
19
+ 0.5,
20
+ 0.5,
21
+ 0.5
22
+ ],
23
+ "input_data_format": null,
24
+ "max_frames": 768,
25
+ "merge_size": 2,
26
+ "min_frames": 4,
27
+ "num_frames": null,
28
+ "pad_size": null,
29
+ "patch_size": 16,
30
+ "processor_class": "Qwen3VLProcessor",
31
+ "resample": 3,
32
+ "rescale_factor": 0.00392156862745098,
33
+ "return_metadata": false,
34
+ "size": {
35
+ "longest_edge": 25165824,
36
+ "shortest_edge": 4096
37
+ },
38
+ "temporal_patch_size": 2,
39
+ "video_metadata": null,
40
+ "video_processor_type": "Qwen3VLVideoProcessor"
41
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff