4 Pulls Updated 6 weeks ago
ffa756aaa620 · 195MB
-
general.architecturestarcoder
-
general.file_typeQ8_0
-
starcoder.attention.head_count12
-
starcoder.attention.head_count_kv1
-
starcoder.attention.layer_norm_epsilon1e-05
-
starcoder.block_count20
-
starcoder.context_length8192
-
starcoder.embedding_length768
-
starcoder.feed_forward_length3072
-
tokenizer.ggml.bos_token_id0
-
tokenizer.ggml.eos_token_id0
-
tokenizer.ggml.merges[Ġ Ġ ĠĠ ĠĠ ĠĠĠĠ ĠĠĠĠ ĠĠ Ġ e r ...]
-
tokenizer.ggml.modelgpt2
-
tokenizer.ggml.prerefact
-
tokenizer.ggml.token_type[3 3 3 3 3 ...]
-
tokenizer.ggml.tokens[<|endoftext|> <fim_prefix> <fim_middle> <fim_suffix> <fim_pad> ...]
-
tokenizer.ggml.unknown_token_id0
-
NameTypeShape
-
position_embd.weightF32[768 8192]
-
token_embd.weightQ8_0[768 49152]
-
blk.0.attn_qkv.biasF32[896]
-
blk.0.attn_qkv.weightQ8_0[768 896]
-
blk.0.attn_output.biasF32[768]
-
blk.0.attn_output.weightQ8_0[768 768]
-
blk.0.attn_norm.biasF32[768]
-
blk.0.attn_norm.weightF32[768]
-
blk.0.ffn_norm.biasF32[768]
-
blk.0.ffn_norm.weightF32[768]
-
blk.0.ffn_up.biasF32[3072]
-
blk.0.ffn_up.weightQ8_0[768 3072]
-
blk.0.ffn_down.biasF32[768]
-
blk.0.ffn_down.weightQ8_0[3072 768]
-
blk.1.attn_qkv.biasF32[896]
-
blk.1.attn_qkv.weightQ8_0[768 896]
-
blk.1.attn_output.biasF32[768]
-
blk.1.attn_output.weightQ8_0[768 768]
-
blk.1.attn_norm.biasF32[768]
-
blk.1.attn_norm.weightF32[768]
-
blk.1.ffn_norm.biasF32[768]
-
blk.1.ffn_norm.weightF32[768]
-
blk.1.ffn_up.biasF32[3072]
-
blk.1.ffn_up.weightQ8_0[768 3072]
-
blk.1.ffn_down.biasF32[768]
-
blk.1.ffn_down.weightQ8_0[3072 768]
-
blk.2.attn_qkv.biasF32[896]
-
blk.2.attn_qkv.weightQ8_0[768 896]
-
blk.2.attn_output.biasF32[768]
-
blk.2.attn_output.weightQ8_0[768 768]
-
blk.2.attn_norm.biasF32[768]
-
blk.2.attn_norm.weightF32[768]
-
blk.2.ffn_norm.biasF32[768]
-
blk.2.ffn_norm.weightF32[768]
-
blk.2.ffn_up.biasF32[3072]
-
blk.2.ffn_up.weightQ8_0[768 3072]
-
blk.2.ffn_down.biasF32[768]
-
blk.2.ffn_down.weightQ8_0[3072 768]
-
blk.3.attn_qkv.biasF32[896]
-
blk.3.attn_qkv.weightQ8_0[768 896]
-
blk.3.attn_output.biasF32[768]
-
blk.3.attn_output.weightQ8_0[768 768]
-
blk.3.attn_norm.biasF32[768]
-
blk.3.attn_norm.weightF32[768]
-
blk.3.ffn_norm.biasF32[768]
-
blk.3.ffn_norm.weightF32[768]
-
blk.3.ffn_up.biasF32[3072]
-
blk.3.ffn_up.weightQ8_0[768 3072]
-
blk.3.ffn_down.biasF32[768]
-
blk.3.ffn_down.weightQ8_0[3072 768]
-
blk.4.attn_qkv.biasF32[896]
-
blk.4.attn_qkv.weightQ8_0[768 896]
-
blk.4.attn_output.biasF32[768]
-
blk.4.attn_output.weightQ8_0[768 768]
-
blk.4.attn_norm.biasF32[768]
-
blk.4.attn_norm.weightF32[768]
-
blk.4.ffn_norm.biasF32[768]
-
blk.4.ffn_norm.weightF32[768]
-
blk.4.ffn_up.biasF32[3072]
-
blk.4.ffn_up.weightQ8_0[768 3072]
-
blk.4.ffn_down.biasF32[768]
-
blk.4.ffn_down.weightQ8_0[3072 768]
-
blk.5.attn_qkv.biasF32[896]
-
blk.5.attn_qkv.weightQ8_0[768 896]
-
blk.5.attn_output.biasF32[768]
-
blk.5.attn_output.weightQ8_0[768 768]
-
blk.5.attn_norm.biasF32[768]
-
blk.5.attn_norm.weightF32[768]
-
blk.5.ffn_norm.biasF32[768]
-
blk.5.ffn_norm.weightF32[768]
-
blk.5.ffn_up.biasF32[3072]
-
blk.5.ffn_up.weightQ8_0[768 3072]
-
blk.5.ffn_down.biasF32[768]
-
blk.5.ffn_down.weightQ8_0[3072 768]
-
blk.6.attn_qkv.biasF32[896]
-
blk.6.attn_qkv.weightQ8_0[768 896]
-
blk.6.attn_output.biasF32[768]
-
blk.6.attn_output.weightQ8_0[768 768]
-
blk.6.attn_norm.biasF32[768]
-
blk.6.attn_norm.weightF32[768]
-
blk.6.ffn_norm.biasF32[768]
-
blk.6.ffn_norm.weightF32[768]
-
blk.6.ffn_up.biasF32[3072]
-
blk.6.ffn_up.weightQ8_0[768 3072]
-
blk.6.ffn_down.biasF32[768]
-
blk.6.ffn_down.weightQ8_0[3072 768]
-
blk.7.attn_qkv.biasF32[896]
-
blk.7.attn_qkv.weightQ8_0[768 896]
-
blk.7.attn_output.biasF32[768]
-
blk.7.attn_output.weightQ8_0[768 768]
-
blk.7.attn_norm.biasF32[768]
-
blk.7.attn_norm.weightF32[768]
-
blk.7.ffn_norm.biasF32[768]
-
blk.7.ffn_norm.weightF32[768]
-
blk.7.ffn_up.biasF32[3072]
-
blk.7.ffn_up.weightQ8_0[768 3072]
-
blk.7.ffn_down.biasF32[768]
-
blk.7.ffn_down.weightQ8_0[3072 768]
-
blk.8.attn_qkv.biasF32[896]
-
blk.8.attn_qkv.weightQ8_0[768 896]
-
blk.8.attn_output.biasF32[768]
-
blk.8.attn_output.weightQ8_0[768 768]
-
blk.8.attn_norm.biasF32[768]
-
blk.8.attn_norm.weightF32[768]
-
blk.8.ffn_norm.biasF32[768]
-
blk.8.ffn_norm.weightF32[768]
-
blk.8.ffn_up.biasF32[3072]
-
blk.8.ffn_up.weightQ8_0[768 3072]
-
blk.8.ffn_down.biasF32[768]
-
blk.8.ffn_down.weightQ8_0[3072 768]
-
blk.9.attn_qkv.biasF32[896]
-
blk.9.attn_qkv.weightQ8_0[768 896]
-
blk.9.attn_output.biasF32[768]
-
blk.9.attn_output.weightQ8_0[768 768]
-
blk.9.attn_norm.biasF32[768]
-
blk.9.attn_norm.weightF32[768]
-
blk.9.ffn_norm.biasF32[768]
-
blk.9.ffn_norm.weightF32[768]
-
blk.9.ffn_up.biasF32[3072]
-
blk.9.ffn_up.weightQ8_0[768 3072]
-
blk.9.ffn_down.biasF32[768]
-
blk.9.ffn_down.weightQ8_0[3072 768]
-
blk.10.attn_qkv.biasF32[896]
-
blk.10.attn_qkv.weightQ8_0[768 896]
-
blk.10.attn_output.biasF32[768]
-
blk.10.attn_output.weightQ8_0[768 768]
-
blk.10.attn_norm.biasF32[768]
-
blk.10.attn_norm.weightF32[768]
-
blk.10.ffn_norm.biasF32[768]
-
blk.10.ffn_norm.weightF32[768]
-
blk.10.ffn_up.biasF32[3072]
-
blk.10.ffn_up.weightQ8_0[768 3072]
-
blk.10.ffn_down.biasF32[768]
-
blk.10.ffn_down.weightQ8_0[3072 768]
-
blk.11.attn_qkv.biasF32[896]
-
blk.11.attn_qkv.weightQ8_0[768 896]
-
blk.11.attn_output.biasF32[768]
-
blk.11.attn_output.weightQ8_0[768 768]
-
blk.11.attn_norm.biasF32[768]
-
blk.11.attn_norm.weightF32[768]
-
blk.11.ffn_norm.biasF32[768]
-
blk.11.ffn_norm.weightF32[768]
-
blk.11.ffn_up.biasF32[3072]
-
blk.11.ffn_up.weightQ8_0[768 3072]
-
blk.11.ffn_down.biasF32[768]
-
blk.11.ffn_down.weightQ8_0[3072 768]
-
blk.12.attn_qkv.biasF32[896]
-
blk.12.attn_qkv.weightQ8_0[768 896]
-
blk.12.attn_output.biasF32[768]
-
blk.12.attn_output.weightQ8_0[768 768]
-
blk.12.attn_norm.biasF32[768]
-
blk.12.attn_norm.weightF32[768]
-
blk.12.ffn_norm.biasF32[768]
-
blk.12.ffn_norm.weightF32[768]
-
blk.12.ffn_up.biasF32[3072]
-
blk.12.ffn_up.weightQ8_0[768 3072]
-
blk.12.ffn_down.biasF32[768]
-
blk.12.ffn_down.weightQ8_0[3072 768]
-
blk.13.attn_qkv.biasF32[896]
-
blk.13.attn_qkv.weightQ8_0[768 896]
-
blk.13.attn_output.biasF32[768]
-
blk.13.attn_output.weightQ8_0[768 768]
-
blk.13.attn_norm.biasF32[768]
-
blk.13.attn_norm.weightF32[768]
-
blk.13.ffn_norm.biasF32[768]
-
blk.13.ffn_norm.weightF32[768]
-
blk.13.ffn_up.biasF32[3072]
-
blk.13.ffn_up.weightQ8_0[768 3072]
-
blk.13.ffn_down.biasF32[768]
-
blk.13.ffn_down.weightQ8_0[3072 768]
-
blk.14.attn_qkv.biasF32[896]
-
blk.14.attn_qkv.weightQ8_0[768 896]
-
blk.14.attn_output.biasF32[768]
-
blk.14.attn_output.weightQ8_0[768 768]
-
blk.14.attn_norm.biasF32[768]
-
blk.14.attn_norm.weightF32[768]
-
blk.14.ffn_norm.biasF32[768]
-
blk.14.ffn_norm.weightF32[768]
-
blk.14.ffn_up.biasF32[3072]
-
blk.14.ffn_up.weightQ8_0[768 3072]
-
blk.14.ffn_down.biasF32[768]
-
blk.14.ffn_down.weightQ8_0[3072 768]
-
blk.15.attn_qkv.biasF32[896]
-
blk.15.attn_qkv.weightQ8_0[768 896]
-
blk.15.attn_output.biasF32[768]
-
blk.15.attn_output.weightQ8_0[768 768]
-
blk.15.attn_norm.biasF32[768]
-
blk.15.attn_norm.weightF32[768]
-
blk.15.ffn_norm.biasF32[768]
-
blk.15.ffn_norm.weightF32[768]
-
blk.15.ffn_up.biasF32[3072]
-
blk.15.ffn_up.weightQ8_0[768 3072]
-
blk.15.ffn_down.biasF32[768]
-
blk.15.ffn_down.weightQ8_0[3072 768]
-
blk.16.attn_qkv.biasF32[896]
-
blk.16.attn_qkv.weightQ8_0[768 896]
-
blk.16.attn_output.biasF32[768]
-
blk.16.attn_output.weightQ8_0[768 768]
-
blk.16.attn_norm.biasF32[768]
-
blk.16.attn_norm.weightF32[768]
-
blk.16.ffn_norm.biasF32[768]
-
blk.16.ffn_norm.weightF32[768]
-
blk.16.ffn_up.biasF32[3072]
-
blk.16.ffn_up.weightQ8_0[768 3072]
-
blk.16.ffn_down.biasF32[768]
-
blk.16.ffn_down.weightQ8_0[3072 768]
-
blk.17.attn_qkv.biasF32[896]
-
blk.17.attn_qkv.weightQ8_0[768 896]
-
blk.17.attn_output.biasF32[768]
-
blk.17.attn_output.weightQ8_0[768 768]
-
blk.17.attn_norm.biasF32[768]
-
blk.17.attn_norm.weightF32[768]
-
blk.17.ffn_norm.biasF32[768]
-
blk.17.ffn_norm.weightF32[768]
-
blk.17.ffn_up.biasF32[3072]
-
blk.17.ffn_up.weightQ8_0[768 3072]
-
blk.17.ffn_down.biasF32[768]
-
blk.17.ffn_down.weightQ8_0[3072 768]
-
blk.18.attn_qkv.biasF32[896]
-
blk.18.attn_qkv.weightQ8_0[768 896]
-
blk.18.attn_output.biasF32[768]
-
blk.18.attn_output.weightQ8_0[768 768]
-
blk.18.attn_norm.biasF32[768]
-
blk.18.attn_norm.weightF32[768]
-
blk.18.ffn_norm.biasF32[768]
-
blk.18.ffn_norm.weightF32[768]
-
blk.18.ffn_up.biasF32[3072]
-
blk.18.ffn_up.weightQ8_0[768 3072]
-
blk.18.ffn_down.biasF32[768]
-
blk.18.ffn_down.weightQ8_0[3072 768]
-
blk.19.attn_qkv.biasF32[896]
-
blk.19.attn_qkv.weightQ8_0[768 896]
-
blk.19.attn_output.biasF32[768]
-
blk.19.attn_output.weightQ8_0[768 768]
-
blk.19.attn_norm.biasF32[768]
-
blk.19.attn_norm.weightF32[768]
-
blk.19.ffn_norm.biasF32[768]
-
blk.19.ffn_norm.weightF32[768]
-
blk.19.ffn_up.biasF32[3072]
-
blk.19.ffn_up.weightQ8_0[768 3072]
-
blk.19.ffn_down.biasF32[768]
-
blk.19.ffn_down.weightQ8_0[3072 768]
-
output_norm.biasF32[768]
-
output_norm.weightF32[768]
Metadata
Tensors
blk.0
blk.1
blk.2
blk.3
blk.4
blk.5
blk.6
blk.7
blk.8
blk.9
blk.10
blk.11
blk.12
blk.13
blk.14
blk.15
blk.16
blk.17
blk.18
blk.19