-
general.architecture
starcoder
-
general.file_type
Q5_1
-
starcoder.attention.head_count
16
-
starcoder.attention.head_count_kv
1
-
starcoder.attention.layer_norm_epsilon
1e-05
-
starcoder.block_count
24
-
starcoder.context_length
8192
-
starcoder.embedding_length
2048
-
starcoder.feed_forward_length
8192
-
tokenizer.ggml.bos_token_id
0
-
tokenizer.ggml.eos_token_id
0
-
tokenizer.ggml.merges
[Ġ Ġ ĠĠ ĠĠ ĠĠĠĠ ĠĠĠĠ ĠĠ Ġ e r ...]
-
tokenizer.ggml.model
gpt2
-
tokenizer.ggml.scores
[0 0 0 0 0 ...]
-
tokenizer.ggml.token_type
[1 1 1 1 1 ...]
-
tokenizer.ggml.tokens
[<|endoftext|> <fim_prefix> <fim_middle> <fim_suffix> <fim_pad> ...]
-
tokenizer.ggml.unknown_token_id
0
-
Name
Type
Shape
-
token_embd.weight
Q5_1
[2048 49152]
-
position_embd.weight
Q5_1
[2048 8192]
-
blk.0.attn_norm.weight
F32
[2048]
-
blk.0.attn_norm.bias
F32
[2048]
-
blk.0.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.0.attn_qkv.bias
F32
[2304]
-
blk.0.attn_output.weight
Q5_1
[2048 2048]
-
blk.0.attn_output.bias
F32
[2048]
-
blk.0.ffn_norm.weight
F32
[2048]
-
blk.0.ffn_norm.bias
F32
[2048]
-
blk.0.ffn_up.weight
Q5_1
[2048 8192]
-
blk.0.ffn_up.bias
F32
[8192]
-
blk.0.ffn_down.weight
Q5_1
[8192 2048]
-
blk.0.ffn_down.bias
F32
[2048]
-
blk.1.attn_norm.weight
F32
[2048]
-
blk.1.attn_norm.bias
F32
[2048]
-
blk.1.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.1.attn_qkv.bias
F32
[2304]
-
blk.1.attn_output.weight
Q5_1
[2048 2048]
-
blk.1.attn_output.bias
F32
[2048]
-
blk.1.ffn_norm.weight
F32
[2048]
-
blk.1.ffn_norm.bias
F32
[2048]
-
blk.1.ffn_up.weight
Q5_1
[2048 8192]
-
blk.1.ffn_up.bias
F32
[8192]
-
blk.1.ffn_down.weight
Q5_1
[8192 2048]
-
blk.1.ffn_down.bias
F32
[2048]
-
blk.2.attn_norm.weight
F32
[2048]
-
blk.2.attn_norm.bias
F32
[2048]
-
blk.2.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.2.attn_qkv.bias
F32
[2304]
-
blk.2.attn_output.weight
Q5_1
[2048 2048]
-
blk.2.attn_output.bias
F32
[2048]
-
blk.2.ffn_norm.weight
F32
[2048]
-
blk.2.ffn_norm.bias
F32
[2048]
-
blk.2.ffn_up.weight
Q5_1
[2048 8192]
-
blk.2.ffn_up.bias
F32
[8192]
-
blk.2.ffn_down.weight
Q5_1
[8192 2048]
-
blk.2.ffn_down.bias
F32
[2048]
-
blk.3.attn_norm.weight
F32
[2048]
-
blk.3.attn_norm.bias
F32
[2048]
-
blk.3.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.3.attn_qkv.bias
F32
[2304]
-
blk.3.attn_output.weight
Q5_1
[2048 2048]
-
blk.3.attn_output.bias
F32
[2048]
-
blk.3.ffn_norm.weight
F32
[2048]
-
blk.3.ffn_norm.bias
F32
[2048]
-
blk.3.ffn_up.weight
Q5_1
[2048 8192]
-
blk.3.ffn_up.bias
F32
[8192]
-
blk.3.ffn_down.weight
Q5_1
[8192 2048]
-
blk.3.ffn_down.bias
F32
[2048]
-
blk.4.attn_norm.weight
F32
[2048]
-
blk.4.attn_norm.bias
F32
[2048]
-
blk.4.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.4.attn_qkv.bias
F32
[2304]
-
blk.4.attn_output.weight
Q5_1
[2048 2048]
-
blk.4.attn_output.bias
F32
[2048]
-
blk.4.ffn_norm.weight
F32
[2048]
-
blk.4.ffn_norm.bias
F32
[2048]
-
blk.4.ffn_up.weight
Q5_1
[2048 8192]
-
blk.4.ffn_up.bias
F32
[8192]
-
blk.4.ffn_down.weight
Q5_1
[8192 2048]
-
blk.4.ffn_down.bias
F32
[2048]
-
blk.5.attn_norm.weight
F32
[2048]
-
blk.5.attn_norm.bias
F32
[2048]
-
blk.5.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.5.attn_qkv.bias
F32
[2304]
-
blk.5.attn_output.weight
Q5_1
[2048 2048]
-
blk.5.attn_output.bias
F32
[2048]
-
blk.5.ffn_norm.weight
F32
[2048]
-
blk.5.ffn_norm.bias
F32
[2048]
-
blk.5.ffn_up.weight
Q5_1
[2048 8192]
-
blk.5.ffn_up.bias
F32
[8192]
-
blk.5.ffn_down.weight
Q5_1
[8192 2048]
-
blk.5.ffn_down.bias
F32
[2048]
-
blk.6.attn_norm.weight
F32
[2048]
-
blk.6.attn_norm.bias
F32
[2048]
-
blk.6.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.6.attn_qkv.bias
F32
[2304]
-
blk.6.attn_output.weight
Q5_1
[2048 2048]
-
blk.6.attn_output.bias
F32
[2048]
-
blk.6.ffn_norm.weight
F32
[2048]
-
blk.6.ffn_norm.bias
F32
[2048]
-
blk.6.ffn_up.weight
Q5_1
[2048 8192]
-
blk.6.ffn_up.bias
F32
[8192]
-
blk.6.ffn_down.weight
Q5_1
[8192 2048]
-
blk.6.ffn_down.bias
F32
[2048]
-
blk.7.attn_norm.weight
F32
[2048]
-
blk.7.attn_norm.bias
F32
[2048]
-
blk.7.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.7.attn_qkv.bias
F32
[2304]
-
blk.7.attn_output.weight
Q5_1
[2048 2048]
-
blk.7.attn_output.bias
F32
[2048]
-
blk.7.ffn_norm.weight
F32
[2048]
-
blk.7.ffn_norm.bias
F32
[2048]
-
blk.7.ffn_up.weight
Q5_1
[2048 8192]
-
blk.7.ffn_up.bias
F32
[8192]
-
blk.7.ffn_down.weight
Q5_1
[8192 2048]
-
blk.7.ffn_down.bias
F32
[2048]
-
blk.8.attn_norm.weight
F32
[2048]
-
blk.8.attn_norm.bias
F32
[2048]
-
blk.8.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.8.attn_qkv.bias
F32
[2304]
-
blk.8.attn_output.weight
Q5_1
[2048 2048]
-
blk.8.attn_output.bias
F32
[2048]
-
blk.8.ffn_norm.weight
F32
[2048]
-
blk.8.ffn_norm.bias
F32
[2048]
-
blk.8.ffn_up.weight
Q5_1
[2048 8192]
-
blk.8.ffn_up.bias
F32
[8192]
-
blk.8.ffn_down.weight
Q5_1
[8192 2048]
-
blk.8.ffn_down.bias
F32
[2048]
-
blk.9.attn_norm.weight
F32
[2048]
-
blk.9.attn_norm.bias
F32
[2048]
-
blk.9.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.9.attn_qkv.bias
F32
[2304]
-
blk.9.attn_output.weight
Q5_1
[2048 2048]
-
blk.9.attn_output.bias
F32
[2048]
-
blk.9.ffn_norm.weight
F32
[2048]
-
blk.9.ffn_norm.bias
F32
[2048]
-
blk.9.ffn_up.weight
Q5_1
[2048 8192]
-
blk.9.ffn_up.bias
F32
[8192]
-
blk.9.ffn_down.weight
Q5_1
[8192 2048]
-
blk.9.ffn_down.bias
F32
[2048]
-
blk.10.attn_norm.weight
F32
[2048]
-
blk.10.attn_norm.bias
F32
[2048]
-
blk.10.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.10.attn_qkv.bias
F32
[2304]
-
blk.10.attn_output.weight
Q5_1
[2048 2048]
-
blk.10.attn_output.bias
F32
[2048]
-
blk.10.ffn_norm.weight
F32
[2048]
-
blk.10.ffn_norm.bias
F32
[2048]
-
blk.10.ffn_up.weight
Q5_1
[2048 8192]
-
blk.10.ffn_up.bias
F32
[8192]
-
blk.10.ffn_down.weight
Q5_1
[8192 2048]
-
blk.10.ffn_down.bias
F32
[2048]
-
blk.11.attn_norm.weight
F32
[2048]
-
blk.11.attn_norm.bias
F32
[2048]
-
blk.11.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.11.attn_qkv.bias
F32
[2304]
-
blk.11.attn_output.weight
Q5_1
[2048 2048]
-
blk.11.attn_output.bias
F32
[2048]
-
blk.11.ffn_norm.weight
F32
[2048]
-
blk.11.ffn_norm.bias
F32
[2048]
-
blk.11.ffn_up.weight
Q5_1
[2048 8192]
-
blk.11.ffn_up.bias
F32
[8192]
-
blk.11.ffn_down.weight
Q5_1
[8192 2048]
-
blk.11.ffn_down.bias
F32
[2048]
-
blk.12.attn_norm.weight
F32
[2048]
-
blk.12.attn_norm.bias
F32
[2048]
-
blk.12.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.12.attn_qkv.bias
F32
[2304]
-
blk.12.attn_output.weight
Q5_1
[2048 2048]
-
blk.12.attn_output.bias
F32
[2048]
-
blk.12.ffn_norm.weight
F32
[2048]
-
blk.12.ffn_norm.bias
F32
[2048]
-
blk.12.ffn_up.weight
Q5_1
[2048 8192]
-
blk.12.ffn_up.bias
F32
[8192]
-
blk.12.ffn_down.weight
Q5_1
[8192 2048]
-
blk.12.ffn_down.bias
F32
[2048]
-
blk.13.attn_norm.weight
F32
[2048]
-
blk.13.attn_norm.bias
F32
[2048]
-
blk.13.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.13.attn_qkv.bias
F32
[2304]
-
blk.13.attn_output.weight
Q5_1
[2048 2048]
-
blk.13.attn_output.bias
F32
[2048]
-
blk.13.ffn_norm.weight
F32
[2048]
-
blk.13.ffn_norm.bias
F32
[2048]
-
blk.13.ffn_up.weight
Q5_1
[2048 8192]
-
blk.13.ffn_up.bias
F32
[8192]
-
blk.13.ffn_down.weight
Q5_1
[8192 2048]
-
blk.13.ffn_down.bias
F32
[2048]
-
blk.14.attn_norm.weight
F32
[2048]
-
blk.14.attn_norm.bias
F32
[2048]
-
blk.14.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.14.attn_qkv.bias
F32
[2304]
-
blk.14.attn_output.weight
Q5_1
[2048 2048]
-
blk.14.attn_output.bias
F32
[2048]
-
blk.14.ffn_norm.weight
F32
[2048]
-
blk.14.ffn_norm.bias
F32
[2048]
-
blk.14.ffn_up.weight
Q5_1
[2048 8192]
-
blk.14.ffn_up.bias
F32
[8192]
-
blk.14.ffn_down.weight
Q5_1
[8192 2048]
-
blk.14.ffn_down.bias
F32
[2048]
-
blk.15.attn_norm.weight
F32
[2048]
-
blk.15.attn_norm.bias
F32
[2048]
-
blk.15.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.15.attn_qkv.bias
F32
[2304]
-
blk.15.attn_output.weight
Q5_1
[2048 2048]
-
blk.15.attn_output.bias
F32
[2048]
-
blk.15.ffn_norm.weight
F32
[2048]
-
blk.15.ffn_norm.bias
F32
[2048]
-
blk.15.ffn_up.weight
Q5_1
[2048 8192]
-
blk.15.ffn_up.bias
F32
[8192]
-
blk.15.ffn_down.weight
Q5_1
[8192 2048]
-
blk.15.ffn_down.bias
F32
[2048]
-
blk.16.attn_norm.weight
F32
[2048]
-
blk.16.attn_norm.bias
F32
[2048]
-
blk.16.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.16.attn_qkv.bias
F32
[2304]
-
blk.16.attn_output.weight
Q5_1
[2048 2048]
-
blk.16.attn_output.bias
F32
[2048]
-
blk.16.ffn_norm.weight
F32
[2048]
-
blk.16.ffn_norm.bias
F32
[2048]
-
blk.16.ffn_up.weight
Q5_1
[2048 8192]
-
blk.16.ffn_up.bias
F32
[8192]
-
blk.16.ffn_down.weight
Q5_1
[8192 2048]
-
blk.16.ffn_down.bias
F32
[2048]
-
blk.17.attn_norm.weight
F32
[2048]
-
blk.17.attn_norm.bias
F32
[2048]
-
blk.17.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.17.attn_qkv.bias
F32
[2304]
-
blk.17.attn_output.weight
Q5_1
[2048 2048]
-
blk.17.attn_output.bias
F32
[2048]
-
blk.17.ffn_norm.weight
F32
[2048]
-
blk.17.ffn_norm.bias
F32
[2048]
-
blk.17.ffn_up.weight
Q5_1
[2048 8192]
-
blk.17.ffn_up.bias
F32
[8192]
-
blk.17.ffn_down.weight
Q5_1
[8192 2048]
-
blk.17.ffn_down.bias
F32
[2048]
-
blk.18.attn_norm.weight
F32
[2048]
-
blk.18.attn_norm.bias
F32
[2048]
-
blk.18.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.18.attn_qkv.bias
F32
[2304]
-
blk.18.attn_output.weight
Q5_1
[2048 2048]
-
blk.18.attn_output.bias
F32
[2048]
-
blk.18.ffn_norm.weight
F32
[2048]
-
blk.18.ffn_norm.bias
F32
[2048]
-
blk.18.ffn_up.weight
Q5_1
[2048 8192]
-
blk.18.ffn_up.bias
F32
[8192]
-
blk.18.ffn_down.weight
Q5_1
[8192 2048]
-
blk.18.ffn_down.bias
F32
[2048]
-
blk.19.attn_norm.weight
F32
[2048]
-
blk.19.attn_norm.bias
F32
[2048]
-
blk.19.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.19.attn_qkv.bias
F32
[2304]
-
blk.19.attn_output.weight
Q5_1
[2048 2048]
-
blk.19.attn_output.bias
F32
[2048]
-
blk.19.ffn_norm.weight
F32
[2048]
-
blk.19.ffn_norm.bias
F32
[2048]
-
blk.19.ffn_up.weight
Q5_1
[2048 8192]
-
blk.19.ffn_up.bias
F32
[8192]
-
blk.19.ffn_down.weight
Q5_1
[8192 2048]
-
blk.19.ffn_down.bias
F32
[2048]
-
blk.20.attn_norm.weight
F32
[2048]
-
blk.20.attn_norm.bias
F32
[2048]
-
blk.20.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.20.attn_qkv.bias
F32
[2304]
-
blk.20.attn_output.weight
Q5_1
[2048 2048]
-
blk.20.attn_output.bias
F32
[2048]
-
blk.20.ffn_norm.weight
F32
[2048]
-
blk.20.ffn_norm.bias
F32
[2048]
-
blk.20.ffn_up.weight
Q5_1
[2048 8192]
-
blk.20.ffn_up.bias
F32
[8192]
-
blk.20.ffn_down.weight
Q5_1
[8192 2048]
-
blk.20.ffn_down.bias
F32
[2048]
-
blk.21.attn_norm.weight
F32
[2048]
-
blk.21.attn_norm.bias
F32
[2048]
-
blk.21.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.21.attn_qkv.bias
F32
[2304]
-
blk.21.attn_output.weight
Q5_1
[2048 2048]
-
blk.21.attn_output.bias
F32
[2048]
-
blk.21.ffn_norm.weight
F32
[2048]
-
blk.21.ffn_norm.bias
F32
[2048]
-
blk.21.ffn_up.weight
Q5_1
[2048 8192]
-
blk.21.ffn_up.bias
F32
[8192]
-
blk.21.ffn_down.weight
Q5_1
[8192 2048]
-
blk.21.ffn_down.bias
F32
[2048]
-
blk.22.attn_norm.weight
F32
[2048]
-
blk.22.attn_norm.bias
F32
[2048]
-
blk.22.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.22.attn_qkv.bias
F32
[2304]
-
blk.22.attn_output.weight
Q5_1
[2048 2048]
-
blk.22.attn_output.bias
F32
[2048]
-
blk.22.ffn_norm.weight
F32
[2048]
-
blk.22.ffn_norm.bias
F32
[2048]
-
blk.22.ffn_up.weight
Q5_1
[2048 8192]
-
blk.22.ffn_up.bias
F32
[8192]
-
blk.22.ffn_down.weight
Q5_1
[8192 2048]
-
blk.22.ffn_down.bias
F32
[2048]
-
blk.23.attn_norm.weight
F32
[2048]
-
blk.23.attn_norm.bias
F32
[2048]
-
blk.23.attn_qkv.weight
Q5_1
[2048 2304]
-
blk.23.attn_qkv.bias
F32
[2304]
-
blk.23.attn_output.weight
Q5_1
[2048 2048]
-
blk.23.attn_output.bias
F32
[2048]
-
blk.23.ffn_norm.weight
F32
[2048]
-
blk.23.ffn_norm.bias
F32
[2048]
-
blk.23.ffn_up.weight
Q5_1
[2048 8192]
-
blk.23.ffn_up.bias
F32
[8192]
-
blk.23.ffn_down.weight
Q5_1
[8192 2048]
-
blk.23.ffn_down.bias
F32
[2048]
-
output_norm.weight
F32
[2048]
-
output_norm.bias
F32
[2048]
-
output.weight
Q6_K
[2048 49152]