-
general.architecture
gpt2
-
general.file_type
Q8_0
-
gpt2.attention.head_count
40
-
gpt2.attention.layer_norm_epsilon
1e-05
-
gpt2.block_count
40
-
gpt2.context_length
2048
-
gpt2.embedding_length
5120
-
gpt2.feed_forward_length
20480
-
tokenizer.ggml.add_bos_token
false
-
tokenizer.ggml.bos_token_id
2
-
tokenizer.ggml.eos_token_id
3
-
tokenizer.ggml.mask_token_id
5
-
tokenizer.ggml.merges
[Ġ Ð Ð ¾ Ð µ Ð ° Ñ Ĥ ...]
-
tokenizer.ggml.model
gpt2
-
tokenizer.ggml.padding_token_id
0
-
tokenizer.ggml.pre
ruGPT-3.5-13B
-
tokenizer.ggml.token_type
[3 3 3 3 1 ...]
-
tokenizer.ggml.tokens
[<pad> <|endoftext|> <s> </s> <unk> ...]
-
tokenizer.ggml.unknown_token_id
1
-
Name
Type
Shape
-
position_embd.weight
F32
[5120 2048]
-
token_embd.weight
Q8_0
[5120 50272]
-
blk.0.attn_qkv.bias
F32
[15360]
-
blk.0.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.0.attn_output.bias
F32
[5120]
-
blk.0.attn_output.weight
Q8_0
[5120 5120]
-
blk.0.attn_norm.bias
F32
[5120]
-
blk.0.attn_norm.weight
F32
[5120]
-
blk.0.ffn_norm.bias
F32
[5120]
-
blk.0.ffn_norm.weight
F32
[5120]
-
blk.0.ffn_up.bias
F32
[20480]
-
blk.0.ffn_up.weight
Q8_0
[5120 20480]
-
blk.0.ffn_down.bias
F32
[5120]
-
blk.0.ffn_down.weight
Q8_0
[20480 5120]
-
blk.1.attn_qkv.bias
F32
[15360]
-
blk.1.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.1.attn_output.bias
F32
[5120]
-
blk.1.attn_output.weight
Q8_0
[5120 5120]
-
blk.1.attn_norm.bias
F32
[5120]
-
blk.1.attn_norm.weight
F32
[5120]
-
blk.1.ffn_norm.bias
F32
[5120]
-
blk.1.ffn_norm.weight
F32
[5120]
-
blk.1.ffn_up.bias
F32
[20480]
-
blk.1.ffn_up.weight
Q8_0
[5120 20480]
-
blk.1.ffn_down.bias
F32
[5120]
-
blk.1.ffn_down.weight
Q8_0
[20480 5120]
-
blk.2.attn_qkv.bias
F32
[15360]
-
blk.2.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.2.attn_output.bias
F32
[5120]
-
blk.2.attn_output.weight
Q8_0
[5120 5120]
-
blk.2.attn_norm.bias
F32
[5120]
-
blk.2.attn_norm.weight
F32
[5120]
-
blk.2.ffn_norm.bias
F32
[5120]
-
blk.2.ffn_norm.weight
F32
[5120]
-
blk.2.ffn_up.bias
F32
[20480]
-
blk.2.ffn_up.weight
Q8_0
[5120 20480]
-
blk.2.ffn_down.bias
F32
[5120]
-
blk.2.ffn_down.weight
Q8_0
[20480 5120]
-
blk.3.attn_norm.bias
F32
[5120]
-
blk.3.attn_norm.weight
F32
[5120]
-
blk.3.attn_qkv.bias
F32
[15360]
-
blk.3.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.3.attn_output.bias
F32
[5120]
-
blk.3.attn_output.weight
Q8_0
[5120 5120]
-
blk.3.ffn_norm.bias
F32
[5120]
-
blk.3.ffn_norm.weight
F32
[5120]
-
blk.3.ffn_up.bias
F32
[20480]
-
blk.3.ffn_up.weight
Q8_0
[5120 20480]
-
blk.3.ffn_down.bias
F32
[5120]
-
blk.3.ffn_down.weight
Q8_0
[20480 5120]
-
blk.4.attn_qkv.bias
F32
[15360]
-
blk.4.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.4.attn_output.bias
F32
[5120]
-
blk.4.attn_output.weight
Q8_0
[5120 5120]
-
blk.4.attn_norm.bias
F32
[5120]
-
blk.4.attn_norm.weight
F32
[5120]
-
blk.4.ffn_norm.bias
F32
[5120]
-
blk.4.ffn_norm.weight
F32
[5120]
-
blk.4.ffn_up.bias
F32
[20480]
-
blk.4.ffn_up.weight
Q8_0
[5120 20480]
-
blk.4.ffn_down.bias
F32
[5120]
-
blk.4.ffn_down.weight
Q8_0
[20480 5120]
-
blk.5.attn_qkv.bias
F32
[15360]
-
blk.5.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.5.attn_output.bias
F32
[5120]
-
blk.5.attn_output.weight
Q8_0
[5120 5120]
-
blk.5.attn_norm.bias
F32
[5120]
-
blk.5.attn_norm.weight
F32
[5120]
-
blk.5.ffn_norm.bias
F32
[5120]
-
blk.5.ffn_norm.weight
F32
[5120]
-
blk.5.ffn_up.bias
F32
[20480]
-
blk.5.ffn_up.weight
Q8_0
[5120 20480]
-
blk.5.ffn_down.bias
F32
[5120]
-
blk.5.ffn_down.weight
Q8_0
[20480 5120]
-
blk.6.attn_qkv.bias
F32
[15360]
-
blk.6.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.6.attn_output.bias
F32
[5120]
-
blk.6.attn_output.weight
Q8_0
[5120 5120]
-
blk.6.attn_norm.bias
F32
[5120]
-
blk.6.attn_norm.weight
F32
[5120]
-
blk.6.ffn_norm.bias
F32
[5120]
-
blk.6.ffn_norm.weight
F32
[5120]
-
blk.6.ffn_up.bias
F32
[20480]
-
blk.6.ffn_up.weight
Q8_0
[5120 20480]
-
blk.6.ffn_down.bias
F32
[5120]
-
blk.6.ffn_down.weight
Q8_0
[20480 5120]
-
blk.7.attn_qkv.bias
F32
[15360]
-
blk.7.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.7.attn_output.bias
F32
[5120]
-
blk.7.attn_output.weight
Q8_0
[5120 5120]
-
blk.7.attn_norm.bias
F32
[5120]
-
blk.7.attn_norm.weight
F32
[5120]
-
blk.7.ffn_norm.bias
F32
[5120]
-
blk.7.ffn_norm.weight
F32
[5120]
-
blk.7.ffn_up.bias
F32
[20480]
-
blk.7.ffn_up.weight
Q8_0
[5120 20480]
-
blk.7.ffn_down.bias
F32
[5120]
-
blk.7.ffn_down.weight
Q8_0
[20480 5120]
-
blk.8.attn_qkv.bias
F32
[15360]
-
blk.8.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.8.attn_output.bias
F32
[5120]
-
blk.8.attn_output.weight
Q8_0
[5120 5120]
-
blk.8.attn_norm.bias
F32
[5120]
-
blk.8.attn_norm.weight
F32
[5120]
-
blk.8.ffn_norm.bias
F32
[5120]
-
blk.8.ffn_norm.weight
F32
[5120]
-
blk.8.ffn_up.bias
F32
[20480]
-
blk.8.ffn_up.weight
Q8_0
[5120 20480]
-
blk.8.ffn_down.bias
F32
[5120]
-
blk.8.ffn_down.weight
Q8_0
[20480 5120]
-
blk.9.attn_qkv.bias
F32
[15360]
-
blk.9.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.9.attn_output.bias
F32
[5120]
-
blk.9.attn_output.weight
Q8_0
[5120 5120]
-
blk.9.attn_norm.bias
F32
[5120]
-
blk.9.attn_norm.weight
F32
[5120]
-
blk.9.ffn_norm.bias
F32
[5120]
-
blk.9.ffn_norm.weight
F32
[5120]
-
blk.9.ffn_up.bias
F32
[20480]
-
blk.9.ffn_up.weight
Q8_0
[5120 20480]
-
blk.9.ffn_down.bias
F32
[5120]
-
blk.9.ffn_down.weight
Q8_0
[20480 5120]
-
blk.10.attn_qkv.bias
F32
[15360]
-
blk.10.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.10.attn_output.bias
F32
[5120]
-
blk.10.attn_output.weight
Q8_0
[5120 5120]
-
blk.10.attn_norm.bias
F32
[5120]
-
blk.10.attn_norm.weight
F32
[5120]
-
blk.10.ffn_norm.bias
F32
[5120]
-
blk.10.ffn_norm.weight
F32
[5120]
-
blk.10.ffn_up.bias
F32
[20480]
-
blk.10.ffn_up.weight
Q8_0
[5120 20480]
-
blk.10.ffn_down.bias
F32
[5120]
-
blk.10.ffn_down.weight
Q8_0
[20480 5120]
-
blk.11.attn_qkv.bias
F32
[15360]
-
blk.11.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.11.attn_output.bias
F32
[5120]
-
blk.11.attn_output.weight
Q8_0
[5120 5120]
-
blk.11.attn_norm.bias
F32
[5120]
-
blk.11.attn_norm.weight
F32
[5120]
-
blk.11.ffn_norm.bias
F32
[5120]
-
blk.11.ffn_norm.weight
F32
[5120]
-
blk.11.ffn_up.bias
F32
[20480]
-
blk.11.ffn_up.weight
Q8_0
[5120 20480]
-
blk.11.ffn_down.bias
F32
[5120]
-
blk.11.ffn_down.weight
Q8_0
[20480 5120]
-
blk.12.attn_qkv.bias
F32
[15360]
-
blk.12.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.12.attn_output.bias
F32
[5120]
-
blk.12.attn_output.weight
Q8_0
[5120 5120]
-
blk.12.attn_norm.bias
F32
[5120]
-
blk.12.attn_norm.weight
F32
[5120]
-
blk.12.ffn_norm.bias
F32
[5120]
-
blk.12.ffn_norm.weight
F32
[5120]
-
blk.12.ffn_up.bias
F32
[20480]
-
blk.12.ffn_up.weight
Q8_0
[5120 20480]
-
blk.12.ffn_down.bias
F32
[5120]
-
blk.12.ffn_down.weight
Q8_0
[20480 5120]
-
blk.13.attn_qkv.bias
F32
[15360]
-
blk.13.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.13.attn_output.bias
F32
[5120]
-
blk.13.attn_output.weight
Q8_0
[5120 5120]
-
blk.13.attn_norm.bias
F32
[5120]
-
blk.13.attn_norm.weight
F32
[5120]
-
blk.13.ffn_norm.bias
F32
[5120]
-
blk.13.ffn_norm.weight
F32
[5120]
-
blk.13.ffn_up.bias
F32
[20480]
-
blk.13.ffn_up.weight
Q8_0
[5120 20480]
-
blk.13.ffn_down.bias
F32
[5120]
-
blk.13.ffn_down.weight
Q8_0
[20480 5120]
-
blk.14.attn_qkv.bias
F32
[15360]
-
blk.14.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.14.attn_norm.bias
F32
[5120]
-
blk.14.attn_norm.weight
F32
[5120]
-
blk.14.attn_output.bias
F32
[5120]
-
blk.14.attn_output.weight
Q8_0
[5120 5120]
-
blk.14.ffn_norm.bias
F32
[5120]
-
blk.14.ffn_norm.weight
F32
[5120]
-
blk.14.ffn_up.bias
F32
[20480]
-
blk.14.ffn_up.weight
Q8_0
[5120 20480]
-
blk.14.ffn_down.bias
F32
[5120]
-
blk.14.ffn_down.weight
Q8_0
[20480 5120]
-
blk.15.attn_qkv.bias
F32
[15360]
-
blk.15.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.15.attn_output.bias
F32
[5120]
-
blk.15.attn_output.weight
Q8_0
[5120 5120]
-
blk.15.attn_norm.bias
F32
[5120]
-
blk.15.attn_norm.weight
F32
[5120]
-
blk.15.ffn_norm.bias
F32
[5120]
-
blk.15.ffn_norm.weight
F32
[5120]
-
blk.15.ffn_up.bias
F32
[20480]
-
blk.15.ffn_up.weight
Q8_0
[5120 20480]
-
blk.15.ffn_down.bias
F32
[5120]
-
blk.15.ffn_down.weight
Q8_0
[20480 5120]
-
blk.16.attn_qkv.bias
F32
[15360]
-
blk.16.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.16.attn_output.bias
F32
[5120]
-
blk.16.attn_output.weight
Q8_0
[5120 5120]
-
blk.16.attn_norm.bias
F32
[5120]
-
blk.16.attn_norm.weight
F32
[5120]
-
blk.16.ffn_norm.bias
F32
[5120]
-
blk.16.ffn_norm.weight
F32
[5120]
-
blk.16.ffn_up.bias
F32
[20480]
-
blk.16.ffn_up.weight
Q8_0
[5120 20480]
-
blk.16.ffn_down.bias
F32
[5120]
-
blk.16.ffn_down.weight
Q8_0
[20480 5120]
-
blk.17.attn_qkv.bias
F32
[15360]
-
blk.17.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.17.attn_output.bias
F32
[5120]
-
blk.17.attn_output.weight
Q8_0
[5120 5120]
-
blk.17.attn_norm.bias
F32
[5120]
-
blk.17.attn_norm.weight
F32
[5120]
-
blk.17.ffn_norm.bias
F32
[5120]
-
blk.17.ffn_norm.weight
F32
[5120]
-
blk.17.ffn_up.bias
F32
[20480]
-
blk.17.ffn_up.weight
Q8_0
[5120 20480]
-
blk.17.ffn_down.bias
F32
[5120]
-
blk.17.ffn_down.weight
Q8_0
[20480 5120]
-
blk.18.attn_norm.bias
F32
[5120]
-
blk.18.attn_norm.weight
F32
[5120]
-
blk.18.attn_qkv.bias
F32
[15360]
-
blk.18.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.18.attn_output.bias
F32
[5120]
-
blk.18.attn_output.weight
Q8_0
[5120 5120]
-
blk.18.ffn_norm.bias
F32
[5120]
-
blk.18.ffn_norm.weight
F32
[5120]
-
blk.18.ffn_up.bias
F32
[20480]
-
blk.18.ffn_up.weight
Q8_0
[5120 20480]
-
blk.18.ffn_down.bias
F32
[5120]
-
blk.18.ffn_down.weight
Q8_0
[20480 5120]
-
blk.19.attn_qkv.bias
F32
[15360]
-
blk.19.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.19.attn_output.bias
F32
[5120]
-
blk.19.attn_output.weight
Q8_0
[5120 5120]
-
blk.19.attn_norm.bias
F32
[5120]
-
blk.19.attn_norm.weight
F32
[5120]
-
blk.19.ffn_norm.bias
F32
[5120]
-
blk.19.ffn_norm.weight
F32
[5120]
-
blk.19.ffn_up.bias
F32
[20480]
-
blk.19.ffn_up.weight
Q8_0
[5120 20480]
-
blk.19.ffn_down.bias
F32
[5120]
-
blk.19.ffn_down.weight
Q8_0
[20480 5120]
-
blk.20.attn_qkv.bias
F32
[15360]
-
blk.20.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.20.attn_output.bias
F32
[5120]
-
blk.20.attn_output.weight
Q8_0
[5120 5120]
-
blk.20.attn_norm.bias
F32
[5120]
-
blk.20.attn_norm.weight
F32
[5120]
-
blk.20.ffn_norm.bias
F32
[5120]
-
blk.20.ffn_norm.weight
F32
[5120]
-
blk.20.ffn_up.bias
F32
[20480]
-
blk.20.ffn_up.weight
Q8_0
[5120 20480]
-
blk.20.ffn_down.bias
F32
[5120]
-
blk.20.ffn_down.weight
Q8_0
[20480 5120]
-
blk.21.attn_qkv.bias
F32
[15360]
-
blk.21.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.21.attn_output.bias
F32
[5120]
-
blk.21.attn_output.weight
Q8_0
[5120 5120]
-
blk.21.attn_norm.bias
F32
[5120]
-
blk.21.attn_norm.weight
F32
[5120]
-
blk.21.ffn_norm.bias
F32
[5120]
-
blk.21.ffn_norm.weight
F32
[5120]
-
blk.21.ffn_up.bias
F32
[20480]
-
blk.21.ffn_up.weight
Q8_0
[5120 20480]
-
blk.21.ffn_down.bias
F32
[5120]
-
blk.21.ffn_down.weight
Q8_0
[20480 5120]
-
blk.22.attn_qkv.bias
F32
[15360]
-
blk.22.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.22.attn_output.bias
F32
[5120]
-
blk.22.attn_output.weight
Q8_0
[5120 5120]
-
blk.22.attn_norm.bias
F32
[5120]
-
blk.22.attn_norm.weight
F32
[5120]
-
blk.22.ffn_norm.bias
F32
[5120]
-
blk.22.ffn_norm.weight
F32
[5120]
-
blk.22.ffn_up.bias
F32
[20480]
-
blk.22.ffn_up.weight
Q8_0
[5120 20480]
-
blk.22.ffn_down.bias
F32
[5120]
-
blk.22.ffn_down.weight
Q8_0
[20480 5120]
-
blk.23.attn_qkv.bias
F32
[15360]
-
blk.23.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.23.attn_output.bias
F32
[5120]
-
blk.23.attn_output.weight
Q8_0
[5120 5120]
-
blk.23.attn_norm.bias
F32
[5120]
-
blk.23.attn_norm.weight
F32
[5120]
-
blk.23.ffn_norm.bias
F32
[5120]
-
blk.23.ffn_norm.weight
F32
[5120]
-
blk.23.ffn_up.bias
F32
[20480]
-
blk.23.ffn_up.weight
Q8_0
[5120 20480]
-
blk.23.ffn_down.bias
F32
[5120]
-
blk.23.ffn_down.weight
Q8_0
[20480 5120]
-
blk.24.attn_qkv.bias
F32
[15360]
-
blk.24.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.24.attn_output.bias
F32
[5120]
-
blk.24.attn_output.weight
Q8_0
[5120 5120]
-
blk.24.attn_norm.bias
F32
[5120]
-
blk.24.attn_norm.weight
F32
[5120]
-
blk.24.ffn_norm.bias
F32
[5120]
-
blk.24.ffn_norm.weight
F32
[5120]
-
blk.24.ffn_up.bias
F32
[20480]
-
blk.24.ffn_up.weight
Q8_0
[5120 20480]
-
blk.24.ffn_down.bias
F32
[5120]
-
blk.24.ffn_down.weight
Q8_0
[20480 5120]
-
blk.25.attn_qkv.bias
F32
[15360]
-
blk.25.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.25.attn_output.bias
F32
[5120]
-
blk.25.attn_output.weight
Q8_0
[5120 5120]
-
blk.25.attn_norm.bias
F32
[5120]
-
blk.25.attn_norm.weight
F32
[5120]
-
blk.25.ffn_norm.bias
F32
[5120]
-
blk.25.ffn_norm.weight
F32
[5120]
-
blk.25.ffn_up.bias
F32
[20480]
-
blk.25.ffn_up.weight
Q8_0
[5120 20480]
-
blk.25.ffn_down.bias
F32
[5120]
-
blk.25.ffn_down.weight
Q8_0
[20480 5120]
-
blk.26.attn_qkv.bias
F32
[15360]
-
blk.26.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.26.attn_output.bias
F32
[5120]
-
blk.26.attn_output.weight
Q8_0
[5120 5120]
-
blk.26.attn_norm.bias
F32
[5120]
-
blk.26.attn_norm.weight
F32
[5120]
-
blk.26.ffn_norm.bias
F32
[5120]
-
blk.26.ffn_norm.weight
F32
[5120]
-
blk.26.ffn_up.bias
F32
[20480]
-
blk.26.ffn_up.weight
Q8_0
[5120 20480]
-
blk.26.ffn_down.bias
F32
[5120]
-
blk.26.ffn_down.weight
Q8_0
[20480 5120]
-
blk.27.attn_qkv.bias
F32
[15360]
-
blk.27.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.27.attn_output.bias
F32
[5120]
-
blk.27.attn_output.weight
Q8_0
[5120 5120]
-
blk.27.attn_norm.bias
F32
[5120]
-
blk.27.attn_norm.weight
F32
[5120]
-
blk.27.ffn_norm.bias
F32
[5120]
-
blk.27.ffn_norm.weight
F32
[5120]
-
blk.27.ffn_up.bias
F32
[20480]
-
blk.27.ffn_up.weight
Q8_0
[5120 20480]
-
blk.27.ffn_down.bias
F32
[5120]
-
blk.27.ffn_down.weight
Q8_0
[20480 5120]
-
blk.28.attn_qkv.bias
F32
[15360]
-
blk.28.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.28.attn_output.bias
F32
[5120]
-
blk.28.attn_output.weight
Q8_0
[5120 5120]
-
blk.28.attn_norm.bias
F32
[5120]
-
blk.28.attn_norm.weight
F32
[5120]
-
blk.28.ffn_norm.bias
F32
[5120]
-
blk.28.ffn_norm.weight
F32
[5120]
-
blk.28.ffn_up.bias
F32
[20480]
-
blk.28.ffn_up.weight
Q8_0
[5120 20480]
-
blk.28.ffn_down.bias
F32
[5120]
-
blk.28.ffn_down.weight
Q8_0
[20480 5120]
-
blk.29.attn_qkv.bias
F32
[15360]
-
blk.29.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.29.attn_norm.bias
F32
[5120]
-
blk.29.attn_norm.weight
F32
[5120]
-
blk.29.attn_output.bias
F32
[5120]
-
blk.29.attn_output.weight
Q8_0
[5120 5120]
-
blk.29.ffn_norm.bias
F32
[5120]
-
blk.29.ffn_norm.weight
F32
[5120]
-
blk.29.ffn_up.bias
F32
[20480]
-
blk.29.ffn_up.weight
Q8_0
[5120 20480]
-
blk.29.ffn_down.bias
F32
[5120]
-
blk.29.ffn_down.weight
Q8_0
[20480 5120]
-
blk.30.attn_qkv.bias
F32
[15360]
-
blk.30.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.30.attn_output.bias
F32
[5120]
-
blk.30.attn_output.weight
Q8_0
[5120 5120]
-
blk.30.attn_norm.bias
F32
[5120]
-
blk.30.attn_norm.weight
F32
[5120]
-
blk.30.ffn_norm.bias
F32
[5120]
-
blk.30.ffn_norm.weight
F32
[5120]
-
blk.30.ffn_up.bias
F32
[20480]
-
blk.30.ffn_up.weight
Q8_0
[5120 20480]
-
blk.30.ffn_down.bias
F32
[5120]
-
blk.30.ffn_down.weight
Q8_0
[20480 5120]
-
blk.31.attn_qkv.bias
F32
[15360]
-
blk.31.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.31.attn_output.bias
F32
[5120]
-
blk.31.attn_output.weight
Q8_0
[5120 5120]
-
blk.31.attn_norm.bias
F32
[5120]
-
blk.31.attn_norm.weight
F32
[5120]
-
blk.31.ffn_norm.bias
F32
[5120]
-
blk.31.ffn_norm.weight
F32
[5120]
-
blk.31.ffn_up.bias
F32
[20480]
-
blk.31.ffn_up.weight
Q8_0
[5120 20480]
-
blk.31.ffn_down.bias
F32
[5120]
-
blk.31.ffn_down.weight
Q8_0
[20480 5120]
-
blk.32.attn_qkv.bias
F32
[15360]
-
blk.32.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.32.attn_output.bias
F32
[5120]
-
blk.32.attn_output.weight
Q8_0
[5120 5120]
-
blk.32.attn_norm.bias
F32
[5120]
-
blk.32.attn_norm.weight
F32
[5120]
-
blk.32.ffn_norm.bias
F32
[5120]
-
blk.32.ffn_norm.weight
F32
[5120]
-
blk.32.ffn_up.bias
F32
[20480]
-
blk.32.ffn_up.weight
Q8_0
[5120 20480]
-
blk.32.ffn_down.bias
F32
[5120]
-
blk.32.ffn_down.weight
Q8_0
[20480 5120]
-
blk.33.attn_norm.bias
F32
[5120]
-
blk.33.attn_norm.weight
F32
[5120]
-
blk.33.attn_qkv.bias
F32
[15360]
-
blk.33.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.33.attn_output.bias
F32
[5120]
-
blk.33.attn_output.weight
Q8_0
[5120 5120]
-
blk.33.ffn_norm.bias
F32
[5120]
-
blk.33.ffn_norm.weight
F32
[5120]
-
blk.33.ffn_up.bias
F32
[20480]
-
blk.33.ffn_up.weight
Q8_0
[5120 20480]
-
blk.33.ffn_down.bias
F32
[5120]
-
blk.33.ffn_down.weight
Q8_0
[20480 5120]
-
blk.34.attn_qkv.bias
F32
[15360]
-
blk.34.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.34.attn_output.bias
F32
[5120]
-
blk.34.attn_output.weight
Q8_0
[5120 5120]
-
blk.34.attn_norm.bias
F32
[5120]
-
blk.34.attn_norm.weight
F32
[5120]
-
blk.34.ffn_norm.bias
F32
[5120]
-
blk.34.ffn_norm.weight
F32
[5120]
-
blk.34.ffn_up.bias
F32
[20480]
-
blk.34.ffn_up.weight
Q8_0
[5120 20480]
-
blk.34.ffn_down.bias
F32
[5120]
-
blk.34.ffn_down.weight
Q8_0
[20480 5120]
-
blk.35.attn_qkv.bias
F32
[15360]
-
blk.35.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.35.attn_output.bias
F32
[5120]
-
blk.35.attn_output.weight
Q8_0
[5120 5120]
-
blk.35.attn_norm.bias
F32
[5120]
-
blk.35.attn_norm.weight
F32
[5120]
-
blk.35.ffn_norm.bias
F32
[5120]
-
blk.35.ffn_norm.weight
F32
[5120]
-
blk.35.ffn_up.bias
F32
[20480]
-
blk.35.ffn_up.weight
Q8_0
[5120 20480]
-
blk.35.ffn_down.bias
F32
[5120]
-
blk.35.ffn_down.weight
Q8_0
[20480 5120]
-
blk.36.attn_qkv.bias
F32
[15360]
-
blk.36.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.36.attn_output.bias
F32
[5120]
-
blk.36.attn_output.weight
Q8_0
[5120 5120]
-
blk.36.attn_norm.bias
F32
[5120]
-
blk.36.attn_norm.weight
F32
[5120]
-
blk.36.ffn_norm.bias
F32
[5120]
-
blk.36.ffn_norm.weight
F32
[5120]
-
blk.36.ffn_up.bias
F32
[20480]
-
blk.36.ffn_up.weight
Q8_0
[5120 20480]
-
blk.36.ffn_down.bias
F32
[5120]
-
blk.36.ffn_down.weight
Q8_0
[20480 5120]
-
blk.37.attn_qkv.bias
F32
[15360]
-
blk.37.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.37.attn_output.bias
F32
[5120]
-
blk.37.attn_output.weight
Q8_0
[5120 5120]
-
blk.37.attn_norm.bias
F32
[5120]
-
blk.37.attn_norm.weight
F32
[5120]
-
blk.37.ffn_norm.bias
F32
[5120]
-
blk.37.ffn_norm.weight
F32
[5120]
-
blk.37.ffn_up.bias
F32
[20480]
-
blk.37.ffn_up.weight
Q8_0
[5120 20480]
-
blk.37.ffn_down.bias
F32
[5120]
-
blk.37.ffn_down.weight
Q8_0
[20480 5120]
-
blk.38.attn_qkv.bias
F32
[15360]
-
blk.38.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.38.attn_output.bias
F32
[5120]
-
blk.38.attn_output.weight
Q8_0
[5120 5120]
-
blk.38.attn_norm.bias
F32
[5120]
-
blk.38.attn_norm.weight
F32
[5120]
-
blk.38.ffn_norm.bias
F32
[5120]
-
blk.38.ffn_norm.weight
F32
[5120]
-
blk.38.ffn_up.bias
F32
[20480]
-
blk.38.ffn_up.weight
Q8_0
[5120 20480]
-
blk.38.ffn_down.bias
F32
[5120]
-
blk.38.ffn_down.weight
Q8_0
[20480 5120]
-
blk.39.attn_qkv.bias
F32
[15360]
-
blk.39.attn_qkv.weight
Q8_0
[5120 15360]
-
blk.39.attn_output.bias
F32
[5120]
-
blk.39.attn_output.weight
Q8_0
[5120 5120]
-
blk.39.attn_norm.bias
F32
[5120]
-
blk.39.attn_norm.weight
F32
[5120]
-
blk.39.ffn_norm.bias
F32
[5120]
-
blk.39.ffn_norm.weight
F32
[5120]
-
blk.39.ffn_up.bias
F32
[20480]
-
blk.39.ffn_up.weight
Q8_0
[5120 20480]
-
blk.39.ffn_down.bias
F32
[5120]
-
blk.39.ffn_down.weight
Q8_0
[20480 5120]
-
output.weight
Q8_0
[5120 50272]
-
output_norm.bias
F32
[5120]
-
output_norm.weight
F32
[5120]