Hugging Face's logo

plaincompute
/
ppocr-mlx

Image-to-Text
MLX
Safetensors
mlx-weights
paddlepaddle-ocr
ppocrv5
ppocrv6
ppdoclayoutv3
pp-structure
apple-silicon
Model card Files
xet
 Community
ppocr-mlx / formula /formula_model_tree.json
jasonni2's picture
jasonni2
Upload folder using huggingface_hub
1e1b9bd verified
Raw
History Blame Contribute Delete
193 kB
{
 "path": "model",
 "class": "PPFormulaNetForConditionalGeneration",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model",
 "class": "PPFormulaNetModel",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder",
 "class": "PPFormulaNetTextModel",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.embed_tokens",
 "class": "PPFormulaNetScaledWordEmbedding",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 50000,
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.embed_positions",
 "class": "PPFormulaNetLearnedPositionalEmbedding",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2562,
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers",
 "class": "ModuleList",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.0",
 "class": "PPFormulaNetDecoderLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.0.self_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.0.self_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.self_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.self_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.self_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.0.activation_fn",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.self_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.encoder_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.0.encoder_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.encoder_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.encoder_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.encoder_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.0.encoder_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.fc1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2048,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2048
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.fc2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 2048
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.0.final_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.1",
 "class": "PPFormulaNetDecoderLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.1.self_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.1.self_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.self_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.self_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.self_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.1.activation_fn",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.self_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.encoder_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.1.encoder_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.encoder_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.encoder_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.encoder_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.1.encoder_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.fc1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2048,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2048
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.fc2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 2048
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.1.final_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.2",
 "class": "PPFormulaNetDecoderLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.2.self_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.2.self_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.self_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.self_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.self_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.2.activation_fn",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.self_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.encoder_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.2.encoder_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.encoder_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.encoder_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.encoder_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.2.encoder_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.fc1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2048,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2048
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.fc2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 2048
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.2.final_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.3",
 "class": "PPFormulaNetDecoderLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.3.self_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.3.self_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.self_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.self_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.self_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.3.activation_fn",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.self_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.encoder_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.3.encoder_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.encoder_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.encoder_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.encoder_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.3.encoder_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.fc1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2048,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2048
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.fc2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 2048
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.3.final_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.4",
 "class": "PPFormulaNetDecoderLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.4.self_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.4.self_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.self_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.self_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.self_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.4.activation_fn",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.self_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.encoder_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.4.encoder_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.encoder_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.encoder_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.encoder_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.4.encoder_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.fc1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2048,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2048
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.fc2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 2048
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.4.final_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.5",
 "class": "PPFormulaNetDecoderLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.5.self_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.5.self_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.self_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.self_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.self_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.5.activation_fn",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.self_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.encoder_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.5.encoder_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.encoder_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.encoder_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.encoder_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.5.encoder_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.fc1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2048,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2048
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.fc2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 2048
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.5.final_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.6",
 "class": "PPFormulaNetDecoderLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.6.self_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.6.self_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.self_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.self_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.self_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.6.activation_fn",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.self_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.encoder_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.6.encoder_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.encoder_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.encoder_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.encoder_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.6.encoder_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.fc1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2048,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2048
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.fc2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 2048
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.6.final_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.7",
 "class": "PPFormulaNetDecoderLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.7.self_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.7.self_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.self_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.self_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.self_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.7.activation_fn",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.self_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.encoder_attn",
 "class": "PPFormulaNetAttention",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.decoder.layers.7.encoder_attn.k_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.encoder_attn.v_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.encoder_attn.q_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.encoder_attn.out_proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.decoder.layers.7.encoder_attn_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.fc1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2048,
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2048
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.fc2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 2048
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layers.7.final_layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.decoder.layernorm_embedding",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.decoder.layer_norm",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder",
 "class": "PPFormulaNetVisionModel",
 "parameters": [
 {
 "name": "pos_embed",
 "shape": [
 1,
 48,
 48,
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.patch_embed",
 "class": "PPFormulaNetPatchEmbeddings",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.patch_embed.projection",
 "class": "Conv2d",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3,
 16,
 16
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers",
 "class": "ModuleList",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.0",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.0.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.0.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.0.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.0.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.0.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.0.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.0.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.0.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.0.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.1",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.1.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.1.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.1.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.1.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.1.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.1.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.1.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.1.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.1.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.2",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.2.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.2.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 95,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 95,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.2.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.2.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.2.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.2.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.2.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.2.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.2.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.3",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.3.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.3.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.3.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.3.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.3.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.3.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.3.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.3.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.3.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.4",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.4.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.4.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.4.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.4.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.4.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.4.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.4.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.4.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.4.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.5",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.5.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.5.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 95,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 95,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.5.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.5.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.5.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.5.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.5.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.5.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.5.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.6",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.6.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.6.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.6.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.6.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.6.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.6.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.6.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.6.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.6.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.7",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.7.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.7.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.7.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.7.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.7.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.7.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.7.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.7.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.7.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.8",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.8.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.8.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 95,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 95,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.8.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.8.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.8.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.8.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.8.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.8.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.8.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.9",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.9.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.9.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.9.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.9.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.9.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.9.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.9.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.9.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.9.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.10",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.10.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.10.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 27,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.10.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.10.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.10.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.10.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.10.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.10.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.10.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.11",
 "class": "PPFormulaNetVisionLayer",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.11.layer_norm1",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.11.attn",
 "class": "PPFormulaNetVisionAttention",
 "parameters": [
 {
 "name": "rel_pos_h",
 "shape": [
 95,
 64
 ],
 "requires_grad": true
 },
 {
 "name": "rel_pos_w",
 "shape": [
 95,
 64
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.11.attn.qkv",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 2304,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 2304
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.11.attn.proj",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.layers.11.layer_norm2",
 "class": "LayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.11.mlp",
 "class": "PPFormulaNetMLPBlock",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.layers.11.mlp.lin1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 3072,
 768
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 3072
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.11.mlp.lin2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 768,
 3072
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 768
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.layers.11.mlp.act",
 "class": "GELUActivation",
 "parameters": [],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 }
 ]
 },
 {
 "path": "model.model.encoder.neck",
 "class": "PPFormulaNetVisionNeck",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.neck.conv1",
 "class": "Conv2d",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 256,
 768,
 1,
 1
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.neck.layer_norm1",
 "class": "PPFormulaNetLayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 256
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 256
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.neck.conv2",
 "class": "Conv2d",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 256,
 256,
 3,
 3
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.neck.layer_norm2",
 "class": "PPFormulaNetLayerNorm",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 256
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 256
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 },
 {
 "path": "model.model.encoder.multi_modal_projector",
 "class": "PPFormulaNetMultiModalProjector",
 "parameters": [],
 "buffers": [],
 "children": [
 {
 "path": "model.model.encoder.multi_modal_projector.conv1",
 "class": "Conv2d",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 256,
 3,
 3
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.multi_modal_projector.conv2",
 "class": "Conv2d",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 1024,
 512,
 3,
 3
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.multi_modal_projector.linear_1",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 1024,
 1024
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 1024
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 },
 {
 "path": "model.model.encoder.multi_modal_projector.linear_2",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 512,
 1024
 ],
 "requires_grad": true
 },
 {
 "name": "bias",
 "shape": [
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
 }
 ]
 }
 ]
 },
 {
 "path": "model.lm_head",
 "class": "Linear",
 "parameters": [
 {
 "name": "weight",
 "shape": [
 50000,
 512
 ],
 "requires_grad": true
 }
 ],
 "buffers": [],
 "children": []
 }
 ]
}