gpt2
bigsmall
compression
lossless
gpt2-bigsmall / bigsmall.index.json
wpferrell's picture
Initial upload: GPT-2 compressed with BigSmall (lossless, 75.53% ratio)
7214371 verified
{
"metadata": {
"bigsmall_version": "1.0.0",
"container_version": 1,
"format": "fp32",
"mode": "balanced",
"model_type": "llm",
"total_size": 413973499,
"total_raw_size": 548090880,
"ratio_pct": 75.53008344163655,
"shard_count": 1,
"tensor_count": 160,
"shards": [
"model.bs"
]
},
"weight_map": {
"h.0.attn.bias": "model.bs",
"h.0.attn.c_attn.bias": "model.bs",
"h.0.attn.c_attn.weight": "model.bs",
"h.0.attn.c_proj.bias": "model.bs",
"h.0.attn.c_proj.weight": "model.bs",
"h.0.ln_1.bias": "model.bs",
"h.0.ln_1.weight": "model.bs",
"h.0.ln_2.bias": "model.bs",
"h.0.ln_2.weight": "model.bs",
"h.0.mlp.c_fc.bias": "model.bs",
"h.0.mlp.c_fc.weight": "model.bs",
"h.0.mlp.c_proj.bias": "model.bs",
"h.0.mlp.c_proj.weight": "model.bs",
"h.1.attn.bias": "model.bs",
"h.1.attn.c_attn.bias": "model.bs",
"h.1.attn.c_attn.weight": "model.bs",
"h.1.attn.c_proj.bias": "model.bs",
"h.1.attn.c_proj.weight": "model.bs",
"h.1.ln_1.bias": "model.bs",
"h.1.ln_1.weight": "model.bs",
"h.1.ln_2.bias": "model.bs",
"h.1.ln_2.weight": "model.bs",
"h.1.mlp.c_fc.bias": "model.bs",
"h.1.mlp.c_fc.weight": "model.bs",
"h.1.mlp.c_proj.bias": "model.bs",
"h.1.mlp.c_proj.weight": "model.bs",
"h.10.attn.bias": "model.bs",
"h.10.attn.c_attn.bias": "model.bs",
"h.10.attn.c_attn.weight": "model.bs",
"h.10.attn.c_proj.bias": "model.bs",
"h.10.attn.c_proj.weight": "model.bs",
"h.10.ln_1.bias": "model.bs",
"h.10.ln_1.weight": "model.bs",
"h.10.ln_2.bias": "model.bs",
"h.10.ln_2.weight": "model.bs",
"h.10.mlp.c_fc.bias": "model.bs",
"h.10.mlp.c_fc.weight": "model.bs",
"h.10.mlp.c_proj.bias": "model.bs",
"h.10.mlp.c_proj.weight": "model.bs",
"h.11.attn.bias": "model.bs",
"h.11.attn.c_attn.bias": "model.bs",
"h.11.attn.c_attn.weight": "model.bs",
"h.11.attn.c_proj.bias": "model.bs",
"h.11.attn.c_proj.weight": "model.bs",
"h.11.ln_1.bias": "model.bs",
"h.11.ln_1.weight": "model.bs",
"h.11.ln_2.bias": "model.bs",
"h.11.ln_2.weight": "model.bs",
"h.11.mlp.c_fc.bias": "model.bs",
"h.11.mlp.c_fc.weight": "model.bs",
"h.11.mlp.c_proj.bias": "model.bs",
"h.11.mlp.c_proj.weight": "model.bs",
"h.2.attn.bias": "model.bs",
"h.2.attn.c_attn.bias": "model.bs",
"h.2.attn.c_attn.weight": "model.bs",
"h.2.attn.c_proj.bias": "model.bs",
"h.2.attn.c_proj.weight": "model.bs",
"h.2.ln_1.bias": "model.bs",
"h.2.ln_1.weight": "model.bs",
"h.2.ln_2.bias": "model.bs",
"h.2.ln_2.weight": "model.bs",
"h.2.mlp.c_fc.bias": "model.bs",
"h.2.mlp.c_fc.weight": "model.bs",
"h.2.mlp.c_proj.bias": "model.bs",
"h.2.mlp.c_proj.weight": "model.bs",
"h.3.attn.bias": "model.bs",
"h.3.attn.c_attn.bias": "model.bs",
"h.3.attn.c_attn.weight": "model.bs",
"h.3.attn.c_proj.bias": "model.bs",
"h.3.attn.c_proj.weight": "model.bs",
"h.3.ln_1.bias": "model.bs",
"h.3.ln_1.weight": "model.bs",
"h.3.ln_2.bias": "model.bs",
"h.3.ln_2.weight": "model.bs",
"h.3.mlp.c_fc.bias": "model.bs",
"h.3.mlp.c_fc.weight": "model.bs",
"h.3.mlp.c_proj.bias": "model.bs",
"h.3.mlp.c_proj.weight": "model.bs",
"h.4.attn.bias": "model.bs",
"h.4.attn.c_attn.bias": "model.bs",
"h.4.attn.c_attn.weight": "model.bs",
"h.4.attn.c_proj.bias": "model.bs",
"h.4.attn.c_proj.weight": "model.bs",
"h.4.ln_1.bias": "model.bs",
"h.4.ln_1.weight": "model.bs",
"h.4.ln_2.bias": "model.bs",
"h.4.ln_2.weight": "model.bs",
"h.4.mlp.c_fc.bias": "model.bs",
"h.4.mlp.c_fc.weight": "model.bs",
"h.4.mlp.c_proj.bias": "model.bs",
"h.4.mlp.c_proj.weight": "model.bs",
"h.5.attn.bias": "model.bs",
"h.5.attn.c_attn.bias": "model.bs",
"h.5.attn.c_attn.weight": "model.bs",
"h.5.attn.c_proj.bias": "model.bs",
"h.5.attn.c_proj.weight": "model.bs",
"h.5.ln_1.bias": "model.bs",
"h.5.ln_1.weight": "model.bs",
"h.5.ln_2.bias": "model.bs",
"h.5.ln_2.weight": "model.bs",
"h.5.mlp.c_fc.bias": "model.bs",
"h.5.mlp.c_fc.weight": "model.bs",
"h.5.mlp.c_proj.bias": "model.bs",
"h.5.mlp.c_proj.weight": "model.bs",
"h.6.attn.bias": "model.bs",
"h.6.attn.c_attn.bias": "model.bs",
"h.6.attn.c_attn.weight": "model.bs",
"h.6.attn.c_proj.bias": "model.bs",
"h.6.attn.c_proj.weight": "model.bs",
"h.6.ln_1.bias": "model.bs",
"h.6.ln_1.weight": "model.bs",
"h.6.ln_2.bias": "model.bs",
"h.6.ln_2.weight": "model.bs",
"h.6.mlp.c_fc.bias": "model.bs",
"h.6.mlp.c_fc.weight": "model.bs",
"h.6.mlp.c_proj.bias": "model.bs",
"h.6.mlp.c_proj.weight": "model.bs",
"h.7.attn.bias": "model.bs",
"h.7.attn.c_attn.bias": "model.bs",
"h.7.attn.c_attn.weight": "model.bs",
"h.7.attn.c_proj.bias": "model.bs",
"h.7.attn.c_proj.weight": "model.bs",
"h.7.ln_1.bias": "model.bs",
"h.7.ln_1.weight": "model.bs",
"h.7.ln_2.bias": "model.bs",
"h.7.ln_2.weight": "model.bs",
"h.7.mlp.c_fc.bias": "model.bs",
"h.7.mlp.c_fc.weight": "model.bs",
"h.7.mlp.c_proj.bias": "model.bs",
"h.7.mlp.c_proj.weight": "model.bs",
"h.8.attn.bias": "model.bs",
"h.8.attn.c_attn.bias": "model.bs",
"h.8.attn.c_attn.weight": "model.bs",
"h.8.attn.c_proj.bias": "model.bs",
"h.8.attn.c_proj.weight": "model.bs",
"h.8.ln_1.bias": "model.bs",
"h.8.ln_1.weight": "model.bs",
"h.8.ln_2.bias": "model.bs",
"h.8.ln_2.weight": "model.bs",
"h.8.mlp.c_fc.bias": "model.bs",
"h.8.mlp.c_fc.weight": "model.bs",
"h.8.mlp.c_proj.bias": "model.bs",
"h.8.mlp.c_proj.weight": "model.bs",
"h.9.attn.bias": "model.bs",
"h.9.attn.c_attn.bias": "model.bs",
"h.9.attn.c_attn.weight": "model.bs",
"h.9.attn.c_proj.bias": "model.bs",
"h.9.attn.c_proj.weight": "model.bs",
"h.9.ln_1.bias": "model.bs",
"h.9.ln_1.weight": "model.bs",
"h.9.ln_2.bias": "model.bs",
"h.9.ln_2.weight": "model.bs",
"h.9.mlp.c_fc.bias": "model.bs",
"h.9.mlp.c_fc.weight": "model.bs",
"h.9.mlp.c_proj.bias": "model.bs",
"h.9.mlp.c_proj.weight": "model.bs",
"ln_f.bias": "model.bs",
"ln_f.weight": "model.bs",
"wpe.weight": "model.bs",
"wte.weight": "model.bs"
}
}