S1-Omni / config.json
ScienceOne-AI's picture
Upload folder using huggingface_hub
835604b verified
Raw
History Blame Contribute Delete
12 kB
{
"architectures": [
"Qwen3VLForConditionalGeneration"
],
"dtype": "bfloat16",
"eos_token_id": 151645,
"hidden_size": 5120,
"image_token_id": 151655,
"model_type": "qwen3_vl",
"pad_token_id": 151643,
"text_config": {
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 151643,
"dtype": "bfloat16",
"eos_token_id": 151645,
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 5120,
"initializer_range": 0.02,
"intermediate_size": 25600,
"max_position_embeddings": 262144,
"model_type": "qwen3_vl_text",
"num_attention_heads": 64,
"num_hidden_layers": 64,
"num_key_value_heads": 8,
"pad_token_id": 151643,
"rms_norm_eps": 1e-06,
"rope_scaling": {
"mrope_interleaved": true,
"mrope_section": [
24,
20,
20
],
"rope_type": "default"
},
"rope_theta": 5000000,
"use_cache": false,
"vocab_size": 151936
},
"tie_word_embeddings": false,
"transformers_version": "4.57.6",
"video_token_id": 151656,
"vision_config": {
"deepstack_visual_indexes": [
8,
16,
24
],
"depth": 27,
"dtype": "bfloat16",
"hidden_act": "gelu_pytorch_tanh",
"hidden_size": 1152,
"in_channels": 3,
"initializer_range": 0.02,
"intermediate_size": 4304,
"model_type": "qwen3_vl",
"num_heads": 16,
"num_position_embeddings": 2304,
"out_hidden_size": 5120,
"pad_token_id": 151643,
"patch_size": 16,
"spatial_merge_size": 2,
"temporal_patch_size": 2
},
"vision_end_token_id": 151653,
"vision_start_token_id": 151652,
"spec2mol_v2": {
"format": "spec2mol_v2_hf_inline_safetensors",
"tensor_prefix": "spec2mol.",
"small_modules": {
"format": "spec2mol_small_modules_v2",
"tensor_prefix": "spec2mol.",
"modules": {
"s1_omni_ir_linear": {
"prefix": "s1_omni_ir_linear.",
"modality": "ir",
"num_classes": 20,
"vocab": [
"alkane",
"alkene",
"alkyne",
"amine",
"imine",
"nitrile",
"alcohol",
"ether",
"haloalkane",
"aldehyde",
"ketone",
"ester",
"amide",
"arene",
"imidazole",
"pyrazole",
"oxazole",
"isoxazole",
"cyclopropane",
"epoxide"
],
"threshold_source": "0.5",
"thresholds": [
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5
],
"build_args": {
"in_dim": 5120,
"proj_dim": 768,
"n_layers": 2,
"n_heads": 8,
"ffn_mult": 2,
"dropout": 0.1,
"num_queries": 4,
"num_classes": 20,
"feat_dim": 0,
"head_dropout": 0.1
},
"hf_prefix": "spec2mol.s1_omni_ir_linear.",
"tensor_count": 46
},
"s1_omni_raman_linear": {
"prefix": "s1_omni_raman_linear.",
"modality": "raman",
"num_classes": 20,
"vocab": [
"alkane",
"alkene",
"alkyne",
"amine",
"imine",
"nitrile",
"alcohol",
"ether",
"haloalkane",
"aldehyde",
"ketone",
"ester",
"amide",
"arene",
"imidazole",
"pyrazole",
"oxazole",
"isoxazole",
"cyclopropane",
"epoxide"
],
"threshold_source": "0.5",
"thresholds": [
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5
],
"build_args": {
"in_dim": 5120,
"proj_dim": 768,
"n_layers": 2,
"n_heads": 8,
"ffn_mult": 2,
"dropout": 0.1,
"num_queries": 4,
"num_classes": 20,
"feat_dim": 0,
"head_dropout": 0.1
},
"hf_prefix": "spec2mol.s1_omni_raman_linear.",
"tensor_count": 46
},
"s1_omni_uv_linear": {
"prefix": "s1_omni_uv_linear.",
"modality": "uv",
"num_classes": 20,
"vocab": [
"alkane",
"alkene",
"alkyne",
"amine",
"imine",
"nitrile",
"alcohol",
"ether",
"haloalkane",
"aldehyde",
"ketone",
"ester",
"amide",
"arene",
"imidazole",
"pyrazole",
"oxazole",
"isoxazole",
"cyclopropane",
"epoxide"
],
"threshold_source": "0.5",
"thresholds": [
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5
],
"build_args": {
"in_dim": 5120,
"proj_dim": 768,
"n_layers": 2,
"n_heads": 8,
"ffn_mult": 2,
"dropout": 0.1,
"num_queries": 4,
"num_classes": 20,
"feat_dim": 0,
"head_dropout": 0.1
},
"hf_prefix": "spec2mol.s1_omni_uv_linear.",
"tensor_count": 46
},
"atomcount": {
"prefix": "atomcount.",
"build_args": {
"in_dim": 256,
"hidden": [
512,
256
],
"dropout": 0.1,
"out_dim": 27
},
"hf_prefix": "spec2mol.atomcount.",
"tensor_count": 8
},
"motif": {
"prefix": "motif.",
"build_args": {
"in_dim": 256,
"num_classes": 20,
"hidden": [
512,
256
],
"dropout": 0.1,
"vocab": [
"alkane",
"alkene",
"alkyne",
"amine",
"imine",
"nitrile",
"alcohol",
"ether",
"haloalkane",
"aldehyde",
"ketone",
"ester",
"amide",
"arene",
"imidazole",
"pyrazole",
"oxazole",
"isoxazole",
"cyclopropane",
"epoxide"
],
"thresholds_default": [
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5,
0.5
],
"thresholds_tuned": [
0.09999999403953552,
0.3499999940395355,
0.14999999105930328,
0.6499999761581421,
0.8999999761581421,
0.44999998807907104,
0.3499999940395355,
0.6499999761581421,
0.699999988079071,
0.29999998211860657,
0.5,
0.949999988079071,
0.949999988079071,
0.949999988079071,
0.949999988079071,
0.699999988079071,
0.8500000238418579,
0.800000011920929,
0.75,
0.699999988079071
],
"thresholds": [
0.09999999403953552,
0.3499999940395355,
0.14999999105930328,
0.6499999761581421,
0.8999999761581421,
0.44999998807907104,
0.3499999940395355,
0.6499999761581421,
0.699999988079071,
0.29999998211860657,
0.5,
0.949999988079071,
0.949999988079071,
0.949999988079071,
0.949999988079071,
0.699999988079071,
0.8500000238418579,
0.800000011920929,
0.75,
0.699999988079071
],
"threshold_source": "thresholds_tuned"
},
"vocab": [
"alkane",
"alkene",
"alkyne",
"amine",
"imine",
"nitrile",
"alcohol",
"ether",
"haloalkane",
"aldehyde",
"ketone",
"ester",
"amide",
"arene",
"imidazole",
"pyrazole",
"oxazole",
"isoxazole",
"cyclopropane",
"epoxide"
],
"thresholds": [
0.09999999403953552,
0.3499999940395355,
0.14999999105930328,
0.6499999761581421,
0.8999999761581421,
0.44999998807907104,
0.3499999940395355,
0.6499999761581421,
0.699999988079071,
0.29999998211860657,
0.5,
0.949999988079071,
0.949999988079071,
0.949999988079071,
0.949999988079071,
0.699999988079071,
0.8500000238418579,
0.800000011920929,
0.75,
0.699999988079071
],
"threshold_source": "thresholds_tuned",
"hf_prefix": "spec2mol.motif.",
"tensor_count": 8
},
"dmt": {
"prefix": "dmt.",
"config_sha256": "058051516ea7471868a028278121f4e213769ddc2835a9a0e2d5c68ddbd89399",
"keep_module_prefix": true,
"hf_prefix": "spec2mol.dmt.",
"tensor_count": 443
}
},
"source": "model.safetensors.index.json",
"shard": "model-00015-of-00021.safetensors"
},
"runtime": {
"module_source": "hf_safetensors",
"index_file": "model.safetensors.index.json",
"tensor_prefix": "spec2mol.",
"predictors": {
"ir": "hf_safetensors:spec2mol.s1_omni_ir_linear.",
"raman": "hf_safetensors:spec2mol.s1_omni_raman_linear.",
"uv": "hf_safetensors:spec2mol.s1_omni_uv_linear."
},
"route_image_order": [
"ir",
"raman",
"uv"
],
"modality_prompts": {
"ir": "请你分析这张IR谱图",
"raman": "请你分析这张Raman谱图",
"uv": "请你分析这张UV谱图"
},
"missing_modality_policy": {
"predictor": "skip_missing",
"spcformer_context": "zero_fill_missing"
},
"motif_fusion": "onehot_intersection",
"spcformer_context_order": [
"uv",
"ir",
"raman"
],
"motif_source": "s1_omni_linear_present_modalities",
"specformer_usage": [
"dmt_context",
"feat256",
"atom_count"
],
"specformer_motif_head_used": false
}
}
}