Download model.json from patrickbdevaney/LLaDA2.2-mini-orin-nano: direct link, hf CLI and curl.
- Browser
- Download file 20.7 kB
-
https://huggingface.co/patrickbdevaney/LLaDA2.2-mini-orin-nano/resolve/main/model.json
- Command line
-
hf download hf://patrickbdevaney/LLaDA2.2-mini-orin-nano/model.json
-
curl -L -o model.json https://huggingface.co/patrickbdevaney/LLaDA2.2-mini-orin-nano/resolve/main/model.json
20.7 kB
| { | |
| "config": { | |
| "num_hidden_layers": 20, | |
| "hidden_size": 2048, | |
| "intermediate_size": 5120, | |
| "moe_intermediate_size": 512, | |
| "num_experts": 256, | |
| "num_experts_per_tok": 8, | |
| "expert_capacity": 48, | |
| "block_size": 32, | |
| "num_attention_heads": 16, | |
| "num_key_value_heads": 4, | |
| "head_dim": 128, | |
| "rotary_dim": 64, | |
| "rope_theta": 3000000, | |
| "rms_norm_eps": 1e-06, | |
| "routed_scaling_factor": 2.5, | |
| "vocab_size": 157184, | |
| "first_k_dense_replace": 1 | |
| }, | |
| "mask_id": 156895, | |
| "nonexpert_bits": 8, | |
| "lm_bits": 8, | |
| "kv_format": "4c/4", | |
| "rotation": { | |
| "residual_seed": 20261024, | |
| "kv_seed": 20261025, | |
| "v_folded": true | |
| }, | |
| "files": { | |
| "nonexpert_L0.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 21463040, | |
| "i8tiles8", | |
| [ | |
| 10240, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 32194560, | |
| 10731520, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 5120 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L1.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L2.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L3.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L4.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L5.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L6.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L7.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L8.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L9.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L10.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L11.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L12.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L13.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L14.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L15.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L16.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L17.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L18.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "nonexpert_L19.bin": { | |
| "qkv": [ | |
| 0, | |
| 6438912, | |
| "i8tiles8", | |
| [ | |
| 3072, | |
| 2048 | |
| ] | |
| ], | |
| "dense": [ | |
| 6438912, | |
| 4292608, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 2048 | |
| ] | |
| ], | |
| "gate_up": [ | |
| 10731520, | |
| 2146304, | |
| "i8tiles8", | |
| [ | |
| 1024, | |
| 2048 | |
| ] | |
| ], | |
| "down": [ | |
| 12877824, | |
| 1073152, | |
| "i8tiles8", | |
| [ | |
| 2048, | |
| 512 | |
| ] | |
| ] | |
| }, | |
| "lm_head.bin": { | |
| "lm_head": [ | |
| 0, | |
| 329457664, | |
| "i8tiles8", | |
| [ | |
| 157184, | |
| 2048 | |
| ] | |
| ] | |
| }, | |
| "small.bin": { | |
| "L0.input_layernorm.weight": [ | |
| 0, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L0.post_attention_layernorm.weight": [ | |
| 8192, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L0.attention.query_layernorm.weight": [ | |
| 16384, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L0.attention.key_layernorm.weight": [ | |
| 20480, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L1.input_layernorm.weight": [ | |
| 24576, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L1.post_attention_layernorm.weight": [ | |
| 32768, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L1.attention.query_layernorm.weight": [ | |
| 40960, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L1.attention.key_layernorm.weight": [ | |
| 45056, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L1.router": [ | |
| 49152, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L1.expert_bias": [ | |
| 1097728, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L2.input_layernorm.weight": [ | |
| 1101824, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L2.post_attention_layernorm.weight": [ | |
| 1110016, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L2.attention.query_layernorm.weight": [ | |
| 1118208, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L2.attention.key_layernorm.weight": [ | |
| 1122304, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L2.router": [ | |
| 1126400, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L2.expert_bias": [ | |
| 2174976, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L3.input_layernorm.weight": [ | |
| 2179072, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L3.post_attention_layernorm.weight": [ | |
| 2187264, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L3.attention.query_layernorm.weight": [ | |
| 2195456, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L3.attention.key_layernorm.weight": [ | |
| 2199552, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L3.router": [ | |
| 2203648, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L3.expert_bias": [ | |
| 3252224, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L4.input_layernorm.weight": [ | |
| 3256320, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L4.post_attention_layernorm.weight": [ | |
| 3264512, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L4.attention.query_layernorm.weight": [ | |
| 3272704, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L4.attention.key_layernorm.weight": [ | |
| 3276800, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L4.router": [ | |
| 3280896, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L4.expert_bias": [ | |
| 4329472, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L5.input_layernorm.weight": [ | |
| 4333568, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L5.post_attention_layernorm.weight": [ | |
| 4341760, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L5.attention.query_layernorm.weight": [ | |
| 4349952, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L5.attention.key_layernorm.weight": [ | |
| 4354048, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L5.router": [ | |
| 4358144, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L5.expert_bias": [ | |
| 5406720, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L6.input_layernorm.weight": [ | |
| 5410816, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L6.post_attention_layernorm.weight": [ | |
| 5419008, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L6.attention.query_layernorm.weight": [ | |
| 5427200, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L6.attention.key_layernorm.weight": [ | |
| 5431296, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L6.router": [ | |
| 5435392, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L6.expert_bias": [ | |
| 6483968, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L7.input_layernorm.weight": [ | |
| 6488064, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L7.post_attention_layernorm.weight": [ | |
| 6496256, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L7.attention.query_layernorm.weight": [ | |
| 6504448, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L7.attention.key_layernorm.weight": [ | |
| 6508544, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L7.router": [ | |
| 6512640, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L7.expert_bias": [ | |
| 7561216, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L8.input_layernorm.weight": [ | |
| 7565312, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L8.post_attention_layernorm.weight": [ | |
| 7573504, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L8.attention.query_layernorm.weight": [ | |
| 7581696, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L8.attention.key_layernorm.weight": [ | |
| 7585792, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L8.router": [ | |
| 7589888, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L8.expert_bias": [ | |
| 8638464, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L9.input_layernorm.weight": [ | |
| 8642560, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L9.post_attention_layernorm.weight": [ | |
| 8650752, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L9.attention.query_layernorm.weight": [ | |
| 8658944, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L9.attention.key_layernorm.weight": [ | |
| 8663040, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L9.router": [ | |
| 8667136, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L9.expert_bias": [ | |
| 9715712, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L10.input_layernorm.weight": [ | |
| 9719808, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L10.post_attention_layernorm.weight": [ | |
| 9728000, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L10.attention.query_layernorm.weight": [ | |
| 9736192, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L10.attention.key_layernorm.weight": [ | |
| 9740288, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L10.router": [ | |
| 9744384, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L10.expert_bias": [ | |
| 10792960, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L11.input_layernorm.weight": [ | |
| 10797056, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L11.post_attention_layernorm.weight": [ | |
| 10805248, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L11.attention.query_layernorm.weight": [ | |
| 10813440, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L11.attention.key_layernorm.weight": [ | |
| 10817536, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L11.router": [ | |
| 10821632, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L11.expert_bias": [ | |
| 11870208, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L12.input_layernorm.weight": [ | |
| 11874304, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L12.post_attention_layernorm.weight": [ | |
| 11882496, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L12.attention.query_layernorm.weight": [ | |
| 11890688, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L12.attention.key_layernorm.weight": [ | |
| 11894784, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L12.router": [ | |
| 11898880, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L12.expert_bias": [ | |
| 12947456, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L13.input_layernorm.weight": [ | |
| 12951552, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L13.post_attention_layernorm.weight": [ | |
| 12959744, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L13.attention.query_layernorm.weight": [ | |
| 12967936, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L13.attention.key_layernorm.weight": [ | |
| 12972032, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L13.router": [ | |
| 12976128, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L13.expert_bias": [ | |
| 14024704, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L14.input_layernorm.weight": [ | |
| 14028800, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L14.post_attention_layernorm.weight": [ | |
| 14036992, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L14.attention.query_layernorm.weight": [ | |
| 14045184, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L14.attention.key_layernorm.weight": [ | |
| 14049280, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L14.router": [ | |
| 14053376, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L14.expert_bias": [ | |
| 15101952, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L15.input_layernorm.weight": [ | |
| 15106048, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L15.post_attention_layernorm.weight": [ | |
| 15114240, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L15.attention.query_layernorm.weight": [ | |
| 15122432, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L15.attention.key_layernorm.weight": [ | |
| 15126528, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L15.router": [ | |
| 15130624, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L15.expert_bias": [ | |
| 16179200, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L16.input_layernorm.weight": [ | |
| 16183296, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L16.post_attention_layernorm.weight": [ | |
| 16191488, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L16.attention.query_layernorm.weight": [ | |
| 16199680, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L16.attention.key_layernorm.weight": [ | |
| 16203776, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L16.router": [ | |
| 16207872, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L16.expert_bias": [ | |
| 17256448, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L17.input_layernorm.weight": [ | |
| 17260544, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L17.post_attention_layernorm.weight": [ | |
| 17268736, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L17.attention.query_layernorm.weight": [ | |
| 17276928, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L17.attention.key_layernorm.weight": [ | |
| 17281024, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L17.router": [ | |
| 17285120, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L17.expert_bias": [ | |
| 18333696, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L18.input_layernorm.weight": [ | |
| 18337792, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L18.post_attention_layernorm.weight": [ | |
| 18345984, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L18.attention.query_layernorm.weight": [ | |
| 18354176, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L18.attention.key_layernorm.weight": [ | |
| 18358272, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L18.router": [ | |
| 18362368, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L18.expert_bias": [ | |
| 19410944, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "L19.input_layernorm.weight": [ | |
| 19415040, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L19.post_attention_layernorm.weight": [ | |
| 19423232, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "L19.attention.query_layernorm.weight": [ | |
| 19431424, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L19.attention.key_layernorm.weight": [ | |
| 19435520, | |
| 512, | |
| "f32", | |
| [ | |
| 128 | |
| ] | |
| ], | |
| "L19.router": [ | |
| 19439616, | |
| 1048576, | |
| "bf16", | |
| [ | |
| 256, | |
| 2048 | |
| ] | |
| ], | |
| "L19.expert_bias": [ | |
| 20488192, | |
| 1024, | |
| "f32", | |
| [ | |
| 256 | |
| ] | |
| ], | |
| "final_norm": [ | |
| 20492288, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ], | |
| "signs": [ | |
| 20500480, | |
| 8192, | |
| "f32", | |
| [ | |
| 2048 | |
| ] | |
| ] | |
| }, | |
| "embed.bin": { | |
| "embed": [ | |
| 0, | |
| 643825664, | |
| "f16", | |
| [ | |
| 157184, | |
| 2048 | |
| ] | |
| ] | |
| } | |
| } | |
| } |