Danaasa's picture
Upload folder using huggingface_hub
656b04b verified
{
"model_type": "solar",
"architectures": [
"SolarForCausalLM"
],
"pre_weights": [
{
"name": "model.embed_tokens.weight",
"is_embed": true,
"output_space": "running_residual"
}
],
"num_layers_config_key": "num_hidden_layers",
"layer_templates": {
"weights": [
{
"name": "model.layers.${layer_index}.input_layernorm.weight",
"input_space": "running_residual"
},
{
"name": "model.layers.${layer_index}.self_attn.k_proj.weight",
"input_space": "running_residual",
"output_space": "attn_qk_${layer_index}",
"head_split": "output",
"is_kq": true
},
{
"name": "model.layers.${layer_index}.self_attn.q_proj.weight",
"input_space": "running_residual",
"output_space": "attn_qk_${layer_index}",
"head_split": "output",
"is_kq": true
},
{
"name": "model.layers.${layer_index}.self_attn.v_proj.weight",
"input_space": "running_residual",
"output_space": "attn_v_${layer_index}",
"head_split": "output"
},
{
"name": "model.layers.${layer_index}.self_attn.o_proj.weight",
"input_space": "attn_v_${layer_index}",
"output_space": "running_residual",
"head_split": "input"
},
{
"name": "model.layers.${layer_index}.post_attention_layernorm.weight",
"input_space": "running_residual"
},
{
"name": "model.layers.${layer_index}.mlp.gate_proj.weight",
"input_space": "running_residual",
"output_space": "up_${layer_index}"
},
{
"name": "model.layers.${layer_index}.mlp.up_proj.weight",
"input_space": "running_residual",
"output_space": "up_${layer_index}"
},
{
"name": "model.layers.${layer_index}.mlp.down_proj.weight",
"input_space": "up_${layer_index}",
"output_space": "running_residual"
}
]
},
"post_weights": [
{
"name": "model.norm.weight",
"input_space": "running_residual"
},
{
"name": "lm_head.weight",
"input_space": "running_residual",
"is_embed": true,
"optional": true,
"tied_names": [
"model.lm_head.weight"
]
}
]
}