latest
20GB
34B parameter decoder-only transformer pretrained on Finnish, English and code.
262 Pulls Updated 6 months ago
9e62217b0780 · 20GB
-
general.architecturebloom
-
general.file_typeQ4_0
-
bloom.attention.head_count56
-
bloom.attention.head_count_kv56
-
bloom.attention.layer_norm_epsilon1e-05
-
bloom.block_count54
-
bloom.context_length7168
-
bloom.embedding_length7168
-
bloom.feed_forward_length28672
-
tokenizer.ggml.bos_token_id1
-
tokenizer.ggml.eos_token_id2
-
tokenizer.ggml.merges[Ġ Ġ i n ĠĠ ĠĠ Ã ¤ Ġ t ...]
-
tokenizer.ggml.modelgpt2
-
tokenizer.ggml.padding_token_id3
-
tokenizer.ggml.token_type[3 3 3 3 3 ...]
-
tokenizer.ggml.tokens[<unk> <s> </s> <pad> <|user|> ...]
-
tokenizer.ggml.unknown_token_id0
-
NameTypeShape
-
token_embd.weightQ4_0[7168 128000]
-
token_embd_norm.biasF32[7168]
-
token_embd_norm.weightF32[7168]
-
blk.0.attn_norm.biasF32[7168]
-
blk.0.attn_norm.weightF32[7168]
-
blk.0.ffn_down.biasF32[7168]
-
blk.0.ffn_down.weightQ4_0[28672 7168]
-
blk.0.ffn_up.biasF32[28672]
-
blk.0.ffn_up.weightQ4_0[7168 28672]
-
blk.0.ffn_norm.biasF32[7168]
-
blk.0.ffn_norm.weightF32[7168]
-
blk.0.attn_output.biasF32[7168]
-
blk.0.attn_output.weightQ4_0[7168 7168]
-
blk.0.attn_qkv.biasF32[21504]
-
blk.0.attn_qkv.weightQ4_0[7168 21504]
-
blk.1.attn_norm.biasF32[7168]
-
blk.1.attn_norm.weightF32[7168]
-
blk.1.ffn_down.biasF32[7168]
-
blk.1.ffn_down.weightQ4_0[28672 7168]
-
blk.1.ffn_up.biasF32[28672]
-
blk.1.ffn_up.weightQ4_0[7168 28672]
-
blk.1.ffn_norm.biasF32[7168]
-
blk.1.ffn_norm.weightF32[7168]
-
blk.1.attn_output.biasF32[7168]
-
blk.1.attn_output.weightQ4_0[7168 7168]
-
blk.1.attn_qkv.biasF32[21504]
-
blk.1.attn_qkv.weightQ4_0[7168 21504]
-
blk.2.attn_norm.biasF32[7168]
-
blk.2.attn_norm.weightF32[7168]
-
blk.2.ffn_norm.biasF32[7168]
-
blk.2.ffn_norm.weightF32[7168]
-
blk.2.attn_output.biasF32[7168]
-
blk.2.attn_output.weightQ4_0[7168 7168]
-
blk.2.attn_qkv.biasF32[21504]
-
blk.2.attn_qkv.weightQ4_0[7168 21504]
-
blk.2.ffn_down.biasF32[7168]
-
blk.2.ffn_down.weightQ4_0[28672 7168]
-
blk.2.ffn_up.biasF32[28672]
-
blk.2.ffn_up.weightQ4_0[7168 28672]
-
blk.3.attn_norm.biasF32[7168]
-
blk.3.attn_norm.weightF32[7168]
-
blk.3.ffn_down.biasF32[7168]
-
blk.3.ffn_down.weightQ4_0[28672 7168]
-
blk.3.ffn_up.biasF32[28672]
-
blk.3.ffn_up.weightQ4_0[7168 28672]
-
blk.3.ffn_norm.biasF32[7168]
-
blk.3.ffn_norm.weightF32[7168]
-
blk.3.attn_output.biasF32[7168]
-
blk.3.attn_output.weightQ4_0[7168 7168]
-
blk.3.attn_qkv.biasF32[21504]
-
blk.3.attn_qkv.weightQ4_0[7168 21504]
-
blk.4.attn_norm.biasF32[7168]
-
blk.4.attn_norm.weightF32[7168]
-
blk.4.ffn_down.biasF32[7168]
-
blk.4.ffn_down.weightQ4_0[28672 7168]
-
blk.4.ffn_up.biasF32[28672]
-
blk.4.ffn_up.weightQ4_0[7168 28672]
-
blk.4.ffn_norm.biasF32[7168]
-
blk.4.ffn_norm.weightF32[7168]
-
blk.4.attn_output.biasF32[7168]
-
blk.4.attn_output.weightQ4_0[7168 7168]
-
blk.4.attn_qkv.biasF32[21504]
-
blk.4.attn_qkv.weightQ4_0[7168 21504]
-
blk.5.attn_norm.biasF32[7168]
-
blk.5.attn_norm.weightF32[7168]
-
blk.5.ffn_down.biasF32[7168]
-
blk.5.ffn_down.weightQ4_0[28672 7168]
-
blk.5.ffn_up.biasF32[28672]
-
blk.5.ffn_up.weightQ4_0[7168 28672]
-
blk.5.ffn_norm.biasF32[7168]
-
blk.5.ffn_norm.weightF32[7168]
-
blk.5.attn_output.biasF32[7168]
-
blk.5.attn_output.weightQ4_0[7168 7168]
-
blk.5.attn_qkv.biasF32[21504]
-
blk.5.attn_qkv.weightQ4_0[7168 21504]
-
blk.6.attn_norm.biasF32[7168]
-
blk.6.attn_norm.weightF32[7168]
-
blk.6.ffn_norm.biasF32[7168]
-
blk.6.ffn_norm.weightF32[7168]
-
blk.6.attn_output.biasF32[7168]
-
blk.6.attn_output.weightQ4_0[7168 7168]
-
blk.6.attn_qkv.biasF32[21504]
-
blk.6.attn_qkv.weightQ4_0[7168 21504]
-
blk.6.ffn_down.biasF32[7168]
-
blk.6.ffn_down.weightQ4_0[28672 7168]
-
blk.6.ffn_up.biasF32[28672]
-
blk.6.ffn_up.weightQ4_0[7168 28672]
-
blk.7.attn_norm.biasF32[7168]
-
blk.7.attn_norm.weightF32[7168]
-
blk.7.ffn_down.biasF32[7168]
-
blk.7.ffn_down.weightQ4_0[28672 7168]
-
blk.7.ffn_up.biasF32[28672]
-
blk.7.ffn_up.weightQ4_0[7168 28672]
-
blk.7.ffn_norm.biasF32[7168]
-
blk.7.ffn_norm.weightF32[7168]
-
blk.7.attn_output.biasF32[7168]
-
blk.7.attn_output.weightQ4_0[7168 7168]
-
blk.7.attn_qkv.biasF32[21504]
-
blk.7.attn_qkv.weightQ4_0[7168 21504]
-
blk.8.attn_norm.biasF32[7168]
-
blk.8.attn_norm.weightF32[7168]
-
blk.8.ffn_down.biasF32[7168]
-
blk.8.ffn_down.weightQ4_0[28672 7168]
-
blk.8.ffn_up.biasF32[28672]
-
blk.8.ffn_up.weightQ4_0[7168 28672]
-
blk.8.ffn_norm.biasF32[7168]
-
blk.8.ffn_norm.weightF32[7168]
-
blk.8.attn_output.biasF32[7168]
-
blk.8.attn_output.weightQ4_0[7168 7168]
-
blk.8.attn_qkv.biasF32[21504]
-
blk.8.attn_qkv.weightQ4_0[7168 21504]
-
blk.9.attn_norm.biasF32[7168]
-
blk.9.attn_norm.weightF32[7168]
-
blk.9.ffn_down.biasF32[7168]
-
blk.9.ffn_down.weightQ4_0[28672 7168]
-
blk.9.ffn_up.biasF32[28672]
-
blk.9.ffn_up.weightQ4_0[7168 28672]
-
blk.9.ffn_norm.biasF32[7168]
-
blk.9.ffn_norm.weightF32[7168]
-
blk.9.attn_output.biasF32[7168]
-
blk.9.attn_output.weightQ4_0[7168 7168]
-
blk.9.attn_qkv.biasF32[21504]
-
blk.9.attn_qkv.weightQ4_0[7168 21504]
-
blk.10.attn_norm.biasF32[7168]
-
blk.10.attn_norm.weightF32[7168]
-
blk.10.ffn_norm.biasF32[7168]
-
blk.10.ffn_norm.weightF32[7168]
-
blk.10.attn_output.biasF32[7168]
-
blk.10.attn_output.weightQ4_0[7168 7168]
-
blk.10.attn_qkv.biasF32[21504]
-
blk.10.attn_qkv.weightQ4_0[7168 21504]
-
blk.10.ffn_down.biasF32[7168]
-
blk.10.ffn_down.weightQ4_0[28672 7168]
-
blk.10.ffn_up.biasF32[28672]
-
blk.10.ffn_up.weightQ4_0[7168 28672]
-
blk.11.attn_norm.biasF32[7168]
-
blk.11.attn_norm.weightF32[7168]
-
blk.11.ffn_down.biasF32[7168]
-
blk.11.ffn_down.weightQ4_0[28672 7168]
-
blk.11.ffn_up.biasF32[28672]
-
blk.11.ffn_up.weightQ4_0[7168 28672]
-
blk.11.ffn_norm.biasF32[7168]
-
blk.11.ffn_norm.weightF32[7168]
-
blk.11.attn_output.biasF32[7168]
-
blk.11.attn_output.weightQ4_0[7168 7168]
-
blk.11.attn_qkv.biasF32[21504]
-
blk.11.attn_qkv.weightQ4_0[7168 21504]
-
blk.12.attn_norm.biasF32[7168]
-
blk.12.attn_norm.weightF32[7168]
-
blk.12.ffn_down.biasF32[7168]
-
blk.12.ffn_down.weightQ4_0[28672 7168]
-
blk.12.ffn_up.biasF32[28672]
-
blk.12.ffn_up.weightQ4_0[7168 28672]
-
blk.12.ffn_norm.biasF32[7168]
-
blk.12.ffn_norm.weightF32[7168]
-
blk.12.attn_output.biasF32[7168]
-
blk.12.attn_output.weightQ4_0[7168 7168]
-
blk.12.attn_qkv.biasF32[21504]
-
blk.12.attn_qkv.weightQ4_0[7168 21504]
-
blk.13.attn_norm.biasF32[7168]
-
blk.13.attn_norm.weightF32[7168]
-
blk.13.ffn_down.biasF32[7168]
-
blk.13.ffn_down.weightQ4_0[28672 7168]
-
blk.13.ffn_up.biasF32[28672]
-
blk.13.ffn_up.weightQ4_0[7168 28672]
-
blk.13.ffn_norm.biasF32[7168]
-
blk.13.ffn_norm.weightF32[7168]
-
blk.13.attn_output.biasF32[7168]
-
blk.13.attn_output.weightQ4_0[7168 7168]
-
blk.13.attn_qkv.biasF32[21504]
-
blk.13.attn_qkv.weightQ4_0[7168 21504]
-
blk.14.attn_norm.biasF32[7168]
-
blk.14.attn_norm.weightF32[7168]
-
blk.14.ffn_norm.biasF32[7168]
-
blk.14.ffn_norm.weightF32[7168]
-
blk.14.attn_output.biasF32[7168]
-
blk.14.attn_output.weightQ4_0[7168 7168]
-
blk.14.attn_qkv.biasF32[21504]
-
blk.14.attn_qkv.weightQ4_0[7168 21504]
-
blk.14.ffn_down.biasF32[7168]
-
blk.14.ffn_down.weightQ4_0[28672 7168]
-
blk.14.ffn_up.biasF32[28672]
-
blk.14.ffn_up.weightQ4_0[7168 28672]
-
blk.15.attn_norm.biasF32[7168]
-
blk.15.attn_norm.weightF32[7168]
-
blk.15.ffn_down.biasF32[7168]
-
blk.15.ffn_down.weightQ4_0[28672 7168]
-
blk.15.ffn_up.biasF32[28672]
-
blk.15.ffn_up.weightQ4_0[7168 28672]
-
blk.15.ffn_norm.biasF32[7168]
-
blk.15.ffn_norm.weightF32[7168]
-
blk.15.attn_output.biasF32[7168]
-
blk.15.attn_output.weightQ4_0[7168 7168]
-
blk.15.attn_qkv.biasF32[21504]
-
blk.15.attn_qkv.weightQ4_0[7168 21504]
-
blk.16.attn_norm.biasF32[7168]
-
blk.16.attn_norm.weightF32[7168]
-
blk.16.ffn_down.biasF32[7168]
-
blk.16.ffn_down.weightQ4_0[28672 7168]
-
blk.16.ffn_up.biasF32[28672]
-
blk.16.ffn_up.weightQ4_0[7168 28672]
-
blk.16.ffn_norm.biasF32[7168]
-
blk.16.ffn_norm.weightF32[7168]
-
blk.16.attn_output.biasF32[7168]
-
blk.16.attn_output.weightQ4_0[7168 7168]
-
blk.16.attn_qkv.biasF32[21504]
-
blk.16.attn_qkv.weightQ4_0[7168 21504]
-
blk.17.attn_norm.biasF32[7168]
-
blk.17.attn_norm.weightF32[7168]
-
blk.17.ffn_down.biasF32[7168]
-
blk.17.ffn_down.weightQ4_0[28672 7168]
-
blk.17.ffn_up.biasF32[28672]
-
blk.17.ffn_up.weightQ4_0[7168 28672]
-
blk.17.ffn_norm.biasF32[7168]
-
blk.17.ffn_norm.weightF32[7168]
-
blk.17.attn_output.biasF32[7168]
-
blk.17.attn_output.weightQ4_0[7168 7168]
-
blk.17.attn_qkv.biasF32[21504]
-
blk.17.attn_qkv.weightQ4_0[7168 21504]
-
blk.18.attn_norm.biasF32[7168]
-
blk.18.attn_norm.weightF32[7168]
-
blk.18.ffn_norm.biasF32[7168]
-
blk.18.ffn_norm.weightF32[7168]
-
blk.18.attn_output.biasF32[7168]
-
blk.18.attn_output.weightQ4_0[7168 7168]
-
blk.18.attn_qkv.biasF32[21504]
-
blk.18.attn_qkv.weightQ4_0[7168 21504]
-
blk.18.ffn_down.biasF32[7168]
-
blk.18.ffn_down.weightQ4_0[28672 7168]
-
blk.18.ffn_up.biasF32[28672]
-
blk.18.ffn_up.weightQ4_0[7168 28672]
-
blk.19.attn_norm.biasF32[7168]
-
blk.19.attn_norm.weightF32[7168]
-
blk.19.ffn_down.biasF32[7168]
-
blk.19.ffn_down.weightQ4_0[28672 7168]
-
blk.19.ffn_up.biasF32[28672]
-
blk.19.ffn_up.weightQ4_0[7168 28672]
-
blk.19.ffn_norm.biasF32[7168]
-
blk.19.ffn_norm.weightF32[7168]
-
blk.19.attn_output.biasF32[7168]
-
blk.19.attn_output.weightQ4_0[7168 7168]
-
blk.19.attn_qkv.biasF32[21504]
-
blk.19.attn_qkv.weightQ4_0[7168 21504]
-
blk.20.attn_norm.biasF32[7168]
-
blk.20.attn_norm.weightF32[7168]
-
blk.20.ffn_down.biasF32[7168]
-
blk.20.ffn_down.weightQ4_0[28672 7168]
-
blk.20.ffn_up.biasF32[28672]
-
blk.20.ffn_up.weightQ4_0[7168 28672]
-
blk.20.ffn_norm.biasF32[7168]
-
blk.20.ffn_norm.weightF32[7168]
-
blk.20.attn_output.biasF32[7168]
-
blk.20.attn_output.weightQ4_0[7168 7168]
-
blk.20.attn_qkv.biasF32[21504]
-
blk.20.attn_qkv.weightQ4_0[7168 21504]
-
blk.21.attn_norm.biasF32[7168]
-
blk.21.attn_norm.weightF32[7168]
-
blk.21.ffn_down.biasF32[7168]
-
blk.21.ffn_down.weightQ4_0[28672 7168]
-
blk.21.ffn_up.biasF32[28672]
-
blk.21.ffn_up.weightQ4_0[7168 28672]
-
blk.21.ffn_norm.biasF32[7168]
-
blk.21.ffn_norm.weightF32[7168]
-
blk.21.attn_output.biasF32[7168]
-
blk.21.attn_output.weightQ4_0[7168 7168]
-
blk.21.attn_qkv.biasF32[21504]
-
blk.21.attn_qkv.weightQ4_0[7168 21504]
-
blk.22.attn_norm.biasF32[7168]
-
blk.22.attn_norm.weightF32[7168]
-
blk.22.ffn_norm.biasF32[7168]
-
blk.22.ffn_norm.weightF32[7168]
-
blk.22.attn_output.biasF32[7168]
-
blk.22.attn_output.weightQ4_0[7168 7168]
-
blk.22.attn_qkv.biasF32[21504]
-
blk.22.attn_qkv.weightQ4_0[7168 21504]
-
blk.22.ffn_down.biasF32[7168]
-
blk.22.ffn_down.weightQ4_0[28672 7168]
-
blk.22.ffn_up.biasF32[28672]
-
blk.22.ffn_up.weightQ4_0[7168 28672]
-
blk.23.attn_norm.biasF32[7168]
-
blk.23.attn_norm.weightF32[7168]
-
blk.23.ffn_down.biasF32[7168]
-
blk.23.ffn_down.weightQ4_0[28672 7168]
-
blk.23.ffn_up.biasF32[28672]
-
blk.23.ffn_up.weightQ4_0[7168 28672]
-
blk.23.ffn_norm.biasF32[7168]
-
blk.23.ffn_norm.weightF32[7168]
-
blk.23.attn_output.biasF32[7168]
-
blk.23.attn_output.weightQ4_0[7168 7168]
-
blk.23.attn_qkv.biasF32[21504]
-
blk.23.attn_qkv.weightQ4_0[7168 21504]
-
blk.24.attn_norm.biasF32[7168]
-
blk.24.attn_norm.weightF32[7168]
-
blk.24.ffn_down.biasF32[7168]
-
blk.24.ffn_down.weightQ4_0[28672 7168]
-
blk.24.ffn_up.biasF32[28672]
-
blk.24.ffn_up.weightQ4_0[7168 28672]
-
blk.24.ffn_norm.biasF32[7168]
-
blk.24.ffn_norm.weightF32[7168]
-
blk.24.attn_output.biasF32[7168]
-
blk.24.attn_output.weightQ4_0[7168 7168]
-
blk.24.attn_qkv.biasF32[21504]
-
blk.24.attn_qkv.weightQ4_0[7168 21504]
-
blk.25.attn_norm.biasF32[7168]
-
blk.25.attn_norm.weightF32[7168]
-
blk.25.ffn_down.biasF32[7168]
-
blk.25.ffn_down.weightQ4_0[28672 7168]
-
blk.25.ffn_up.biasF32[28672]
-
blk.25.ffn_up.weightQ4_0[7168 28672]
-
blk.25.ffn_norm.biasF32[7168]
-
blk.25.ffn_norm.weightF32[7168]
-
blk.25.attn_output.biasF32[7168]
-
blk.25.attn_output.weightQ4_0[7168 7168]
-
blk.25.attn_qkv.biasF32[21504]
-
blk.25.attn_qkv.weightQ4_0[7168 21504]
-
blk.26.attn_norm.biasF32[7168]
-
blk.26.attn_norm.weightF32[7168]
-
blk.26.ffn_norm.biasF32[7168]
-
blk.26.ffn_norm.weightF32[7168]
-
blk.26.attn_output.biasF32[7168]
-
blk.26.attn_output.weightQ4_0[7168 7168]
-
blk.26.attn_qkv.biasF32[21504]
-
blk.26.attn_qkv.weightQ4_0[7168 21504]
-
blk.26.ffn_down.biasF32[7168]
-
blk.26.ffn_down.weightQ4_0[28672 7168]
-
blk.26.ffn_up.biasF32[28672]
-
blk.26.ffn_up.weightQ4_0[7168 28672]
-
blk.27.attn_norm.biasF32[7168]
-
blk.27.attn_norm.weightF32[7168]
-
blk.27.ffn_down.biasF32[7168]
-
blk.27.ffn_down.weightQ4_0[28672 7168]
-
blk.27.ffn_up.biasF32[28672]
-
blk.27.ffn_up.weightQ4_0[7168 28672]
-
blk.27.ffn_norm.biasF32[7168]
-
blk.27.ffn_norm.weightF32[7168]
-
blk.27.attn_output.biasF32[7168]
-
blk.27.attn_output.weightQ4_0[7168 7168]
-
blk.27.attn_qkv.biasF32[21504]
-
blk.27.attn_qkv.weightQ4_0[7168 21504]
-
blk.28.attn_norm.biasF32[7168]
-
blk.28.attn_norm.weightF32[7168]
-
blk.28.ffn_down.biasF32[7168]
-
blk.28.ffn_down.weightQ4_0[28672 7168]
-
blk.28.ffn_up.biasF32[28672]
-
blk.28.ffn_up.weightQ4_0[7168 28672]
-
blk.28.ffn_norm.biasF32[7168]
-
blk.28.ffn_norm.weightF32[7168]
-
blk.28.attn_output.biasF32[7168]
-
blk.28.attn_output.weightQ4_0[7168 7168]
-
blk.28.attn_qkv.biasF32[21504]
-
blk.28.attn_qkv.weightQ4_0[7168 21504]
-
blk.29.attn_norm.biasF32[7168]
-
blk.29.attn_norm.weightF32[7168]
-
blk.29.ffn_down.biasF32[7168]
-
blk.29.ffn_down.weightQ4_0[28672 7168]
-
blk.29.ffn_up.biasF32[28672]
-
blk.29.ffn_up.weightQ4_0[7168 28672]
-
blk.29.ffn_norm.biasF32[7168]
-
blk.29.ffn_norm.weightF32[7168]
-
blk.29.attn_output.biasF32[7168]
-
blk.29.attn_output.weightQ4_0[7168 7168]
-
blk.29.attn_qkv.biasF32[21504]
-
blk.29.attn_qkv.weightQ4_0[7168 21504]
-
blk.30.attn_norm.biasF32[7168]
-
blk.30.attn_norm.weightF32[7168]
-
blk.30.ffn_norm.biasF32[7168]
-
blk.30.ffn_norm.weightF32[7168]
-
blk.30.attn_output.biasF32[7168]
-
blk.30.attn_output.weightQ4_0[7168 7168]
-
blk.30.attn_qkv.biasF32[21504]
-
blk.30.attn_qkv.weightQ4_0[7168 21504]
-
blk.30.ffn_down.biasF32[7168]
-
blk.30.ffn_down.weightQ4_0[28672 7168]
-
blk.30.ffn_up.biasF32[28672]
-
blk.30.ffn_up.weightQ4_0[7168 28672]
-
blk.31.attn_norm.biasF32[7168]
-
blk.31.attn_norm.weightF32[7168]
-
blk.31.ffn_down.biasF32[7168]
-
blk.31.ffn_down.weightQ4_0[28672 7168]
-
blk.31.ffn_up.biasF32[28672]
-
blk.31.ffn_up.weightQ4_0[7168 28672]
-
blk.31.ffn_norm.biasF32[7168]
-
blk.31.ffn_norm.weightF32[7168]
-
blk.31.attn_output.biasF32[7168]
-
blk.31.attn_output.weightQ4_0[7168 7168]
-
blk.31.attn_qkv.biasF32[21504]
-
blk.31.attn_qkv.weightQ4_0[7168 21504]
-
blk.32.attn_norm.biasF32[7168]
-
blk.32.attn_norm.weightF32[7168]
-
blk.32.ffn_down.biasF32[7168]
-
blk.32.ffn_down.weightQ4_0[28672 7168]
-
blk.32.ffn_up.biasF32[28672]
-
blk.32.ffn_up.weightQ4_0[7168 28672]
-
blk.32.ffn_norm.biasF32[7168]
-
blk.32.ffn_norm.weightF32[7168]
-
blk.32.attn_output.biasF32[7168]
-
blk.32.attn_output.weightQ4_0[7168 7168]
-
blk.32.attn_qkv.biasF32[21504]
-
blk.32.attn_qkv.weightQ4_0[7168 21504]
-
blk.33.attn_norm.biasF32[7168]
-
blk.33.attn_norm.weightF32[7168]
-
blk.33.ffn_down.biasF32[7168]
-
blk.33.ffn_down.weightQ4_0[28672 7168]
-
blk.33.ffn_up.biasF32[28672]
-
blk.33.ffn_up.weightQ4_0[7168 28672]
-
blk.33.ffn_norm.biasF32[7168]
-
blk.33.ffn_norm.weightF32[7168]
-
blk.33.attn_output.biasF32[7168]
-
blk.33.attn_output.weightQ4_0[7168 7168]
-
blk.33.attn_qkv.biasF32[21504]
-
blk.33.attn_qkv.weightQ4_0[7168 21504]
-
blk.34.attn_norm.biasF32[7168]
-
blk.34.attn_norm.weightF32[7168]
-
blk.34.ffn_norm.biasF32[7168]
-
blk.34.ffn_norm.weightF32[7168]
-
blk.34.attn_output.biasF32[7168]
-
blk.34.attn_output.weightQ4_0[7168 7168]
-
blk.34.attn_qkv.biasF32[21504]
-
blk.34.attn_qkv.weightQ4_0[7168 21504]
-
blk.34.ffn_down.biasF32[7168]
-
blk.34.ffn_down.weightQ4_0[28672 7168]
-
blk.34.ffn_up.biasF32[28672]
-
blk.34.ffn_up.weightQ4_0[7168 28672]
-
blk.35.attn_norm.biasF32[7168]
-
blk.35.attn_norm.weightF32[7168]
-
blk.35.ffn_down.biasF32[7168]
-
blk.35.ffn_down.weightQ4_0[28672 7168]
-
blk.35.ffn_up.biasF32[28672]
-
blk.35.ffn_up.weightQ4_0[7168 28672]
-
blk.35.ffn_norm.biasF32[7168]
-
blk.35.ffn_norm.weightF32[7168]
-
blk.35.attn_output.biasF32[7168]
-
blk.35.attn_output.weightQ4_0[7168 7168]
-
blk.35.attn_qkv.biasF32[21504]
-
blk.35.attn_qkv.weightQ4_0[7168 21504]
-
blk.36.attn_norm.biasF32[7168]
-
blk.36.attn_norm.weightF32[7168]
-
blk.36.ffn_down.biasF32[7168]
-
blk.36.ffn_down.weightQ4_0[28672 7168]
-
blk.36.ffn_up.biasF32[28672]
-
blk.36.ffn_up.weightQ4_0[7168 28672]
-
blk.36.ffn_norm.biasF32[7168]
-
blk.36.ffn_norm.weightF32[7168]
-
blk.36.attn_output.biasF32[7168]
-
blk.36.attn_output.weightQ4_0[7168 7168]
-
blk.36.attn_qkv.biasF32[21504]
-
blk.36.attn_qkv.weightQ4_0[7168 21504]
-
blk.37.attn_norm.biasF32[7168]
-
blk.37.attn_norm.weightF32[7168]
-
blk.37.ffn_down.biasF32[7168]
-
blk.37.ffn_down.weightQ4_0[28672 7168]
-
blk.37.ffn_up.biasF32[28672]
-
blk.37.ffn_up.weightQ4_0[7168 28672]
-
blk.37.ffn_norm.biasF32[7168]
-
blk.37.ffn_norm.weightF32[7168]
-
blk.37.attn_output.biasF32[7168]
-
blk.37.attn_output.weightQ4_0[7168 7168]
-
blk.37.attn_qkv.biasF32[21504]
-
blk.37.attn_qkv.weightQ4_0[7168 21504]
-
blk.38.attn_norm.biasF32[7168]
-
blk.38.attn_norm.weightF32[7168]
-
blk.38.ffn_norm.biasF32[7168]
-
blk.38.ffn_norm.weightF32[7168]
-
blk.38.attn_output.biasF32[7168]
-
blk.38.attn_output.weightQ4_0[7168 7168]
-
blk.38.attn_qkv.biasF32[21504]
-
blk.38.attn_qkv.weightQ4_0[7168 21504]
-
blk.38.ffn_down.biasF32[7168]
-
blk.38.ffn_down.weightQ4_0[28672 7168]
-
blk.38.ffn_up.biasF32[28672]
-
blk.38.ffn_up.weightQ4_0[7168 28672]
-
blk.39.attn_norm.biasF32[7168]
-
blk.39.attn_norm.weightF32[7168]
-
blk.39.ffn_down.biasF32[7168]
-
blk.39.ffn_down.weightQ4_0[28672 7168]
-
blk.39.ffn_up.biasF32[28672]
-
blk.39.ffn_up.weightQ4_0[7168 28672]
-
blk.39.ffn_norm.biasF32[7168]
-
blk.39.ffn_norm.weightF32[7168]
-
blk.39.attn_output.biasF32[7168]
-
blk.39.attn_output.weightQ4_0[7168 7168]
-
blk.39.attn_qkv.biasF32[21504]
-
blk.39.attn_qkv.weightQ4_0[7168 21504]
-
blk.40.attn_norm.biasF32[7168]
-
blk.40.attn_norm.weightF32[7168]
-
blk.40.ffn_down.biasF32[7168]
-
blk.40.ffn_down.weightQ4_0[28672 7168]
-
blk.40.ffn_up.biasF32[28672]
-
blk.40.ffn_up.weightQ4_0[7168 28672]
-
blk.40.ffn_norm.biasF32[7168]
-
blk.40.ffn_norm.weightF32[7168]
-
blk.40.attn_output.biasF32[7168]
-
blk.40.attn_output.weightQ4_0[7168 7168]
-
blk.40.attn_qkv.biasF32[21504]
-
blk.40.attn_qkv.weightQ4_0[7168 21504]
-
blk.41.attn_norm.biasF32[7168]
-
blk.41.attn_norm.weightF32[7168]
-
blk.41.ffn_down.biasF32[7168]
-
blk.41.ffn_down.weightQ4_0[28672 7168]
-
blk.41.ffn_up.biasF32[28672]
-
blk.41.ffn_up.weightQ4_0[7168 28672]
-
blk.41.ffn_norm.biasF32[7168]
-
blk.41.ffn_norm.weightF32[7168]
-
blk.41.attn_output.biasF32[7168]
-
blk.41.attn_output.weightQ4_0[7168 7168]
-
blk.41.attn_qkv.biasF32[21504]
-
blk.41.attn_qkv.weightQ4_0[7168 21504]
-
blk.42.attn_norm.biasF32[7168]
-
blk.42.attn_norm.weightF32[7168]
-
blk.42.ffn_norm.biasF32[7168]
-
blk.42.ffn_norm.weightF32[7168]
-
blk.42.attn_output.biasF32[7168]
-
blk.42.attn_output.weightQ4_0[7168 7168]
-
blk.42.attn_qkv.biasF32[21504]
-
blk.42.attn_qkv.weightQ4_0[7168 21504]
-
blk.42.ffn_down.biasF32[7168]
-
blk.42.ffn_down.weightQ4_0[28672 7168]
-
blk.42.ffn_up.biasF32[28672]
-
blk.42.ffn_up.weightQ4_0[7168 28672]
-
blk.43.attn_norm.biasF32[7168]
-
blk.43.attn_norm.weightF32[7168]
-
blk.43.ffn_down.biasF32[7168]
-
blk.43.ffn_down.weightQ4_0[28672 7168]
-
blk.43.ffn_up.biasF32[28672]
-
blk.43.ffn_up.weightQ4_0[7168 28672]
-
blk.43.ffn_norm.biasF32[7168]
-
blk.43.ffn_norm.weightF32[7168]
-
blk.43.attn_output.biasF32[7168]
-
blk.43.attn_output.weightQ4_0[7168 7168]
-
blk.43.attn_qkv.biasF32[21504]
-
blk.43.attn_qkv.weightQ4_0[7168 21504]
-
blk.44.attn_norm.biasF32[7168]
-
blk.44.attn_norm.weightF32[7168]
-
blk.44.ffn_down.biasF32[7168]
-
blk.44.ffn_down.weightQ4_0[28672 7168]
-
blk.44.ffn_up.biasF32[28672]
-
blk.44.ffn_up.weightQ4_0[7168 28672]
-
blk.44.ffn_norm.biasF32[7168]
-
blk.44.ffn_norm.weightF32[7168]
-
blk.44.attn_output.biasF32[7168]
-
blk.44.attn_output.weightQ4_0[7168 7168]
-
blk.44.attn_qkv.biasF32[21504]
-
blk.44.attn_qkv.weightQ4_0[7168 21504]
-
blk.45.attn_norm.biasF32[7168]
-
blk.45.attn_norm.weightF32[7168]
-
blk.45.ffn_down.biasF32[7168]
-
blk.45.ffn_down.weightQ4_0[28672 7168]
-
blk.45.ffn_up.biasF32[28672]
-
blk.45.ffn_up.weightQ4_0[7168 28672]
-
blk.45.ffn_norm.biasF32[7168]
-
blk.45.ffn_norm.weightF32[7168]
-
blk.45.attn_output.biasF32[7168]
-
blk.45.attn_output.weightQ4_0[7168 7168]
-
blk.45.attn_qkv.biasF32[21504]
-
blk.45.attn_qkv.weightQ4_0[7168 21504]
-
blk.46.attn_norm.biasF32[7168]
-
blk.46.attn_norm.weightF32[7168]
-
blk.46.ffn_norm.biasF32[7168]
-
blk.46.ffn_norm.weightF32[7168]
-
blk.46.attn_output.biasF32[7168]
-
blk.46.attn_output.weightQ4_0[7168 7168]
-
blk.46.attn_qkv.biasF32[21504]
-
blk.46.attn_qkv.weightQ4_0[7168 21504]
-
blk.46.ffn_down.biasF32[7168]
-
blk.46.ffn_down.weightQ4_0[28672 7168]
-
blk.46.ffn_up.biasF32[28672]
-
blk.46.ffn_up.weightQ4_0[7168 28672]
-
blk.47.attn_norm.biasF32[7168]
-
blk.47.attn_norm.weightF32[7168]
-
blk.47.ffn_down.biasF32[7168]
-
blk.47.ffn_down.weightQ4_0[28672 7168]
-
blk.47.ffn_up.biasF32[28672]
-
blk.47.ffn_up.weightQ4_0[7168 28672]
-
blk.47.ffn_norm.biasF32[7168]
-
blk.47.ffn_norm.weightF32[7168]
-
blk.47.attn_output.biasF32[7168]
-
blk.47.attn_output.weightQ4_0[7168 7168]
-
blk.47.attn_qkv.biasF32[21504]
-
blk.47.attn_qkv.weightQ4_0[7168 21504]
-
blk.48.attn_norm.biasF32[7168]
-
blk.48.attn_norm.weightF32[7168]
-
blk.48.ffn_down.biasF32[7168]
-
blk.48.ffn_down.weightQ4_0[28672 7168]
-
blk.48.ffn_up.biasF32[28672]
-
blk.48.ffn_up.weightQ4_0[7168 28672]
-
blk.48.ffn_norm.biasF32[7168]
-
blk.48.ffn_norm.weightF32[7168]
-
blk.48.attn_output.biasF32[7168]
-
blk.48.attn_output.weightQ4_0[7168 7168]
-
blk.48.attn_qkv.biasF32[21504]
-
blk.48.attn_qkv.weightQ4_0[7168 21504]
-
blk.49.attn_norm.biasF32[7168]
-
blk.49.attn_norm.weightF32[7168]
-
blk.49.ffn_down.biasF32[7168]
-
blk.49.ffn_down.weightQ4_0[28672 7168]
-
blk.49.ffn_up.biasF32[28672]
-
blk.49.ffn_up.weightQ4_0[7168 28672]
-
blk.49.ffn_norm.biasF32[7168]
-
blk.49.ffn_norm.weightF32[7168]
-
blk.49.attn_output.biasF32[7168]
-
blk.49.attn_output.weightQ4_0[7168 7168]
-
blk.49.attn_qkv.biasF32[21504]
-
blk.49.attn_qkv.weightQ4_0[7168 21504]
-
blk.50.attn_norm.biasF32[7168]
-
blk.50.attn_norm.weightF32[7168]
-
blk.50.ffn_norm.biasF32[7168]
-
blk.50.ffn_norm.weightF32[7168]
-
blk.50.attn_output.biasF32[7168]
-
blk.50.attn_output.weightQ4_0[7168 7168]
-
blk.50.attn_qkv.biasF32[21504]
-
blk.50.attn_qkv.weightQ4_0[7168 21504]
-
blk.50.ffn_down.biasF32[7168]
-
blk.50.ffn_down.weightQ4_0[28672 7168]
-
blk.50.ffn_up.biasF32[28672]
-
blk.50.ffn_up.weightQ4_0[7168 28672]
-
blk.51.attn_norm.biasF32[7168]
-
blk.51.attn_norm.weightF32[7168]
-
blk.51.ffn_down.biasF32[7168]
-
blk.51.ffn_down.weightQ4_0[28672 7168]
-
blk.51.ffn_up.biasF32[28672]
-
blk.51.ffn_up.weightQ4_0[7168 28672]
-
blk.51.ffn_norm.biasF32[7168]
-
blk.51.ffn_norm.weightF32[7168]
-
blk.51.attn_output.biasF32[7168]
-
blk.51.attn_output.weightQ4_0[7168 7168]
-
blk.51.attn_qkv.biasF32[21504]
-
blk.51.attn_qkv.weightQ4_0[7168 21504]
-
blk.52.attn_norm.biasF32[7168]
-
blk.52.attn_norm.weightF32[7168]
-
blk.52.ffn_down.biasF32[7168]
-
blk.52.ffn_down.weightQ4_0[28672 7168]
-
blk.52.ffn_up.biasF32[28672]
-
blk.52.ffn_up.weightQ4_0[7168 28672]
-
blk.52.ffn_norm.biasF32[7168]
-
blk.52.ffn_norm.weightF32[7168]
-
blk.52.attn_output.biasF32[7168]
-
blk.52.attn_output.weightQ4_0[7168 7168]
-
blk.52.attn_qkv.biasF32[21504]
-
blk.52.attn_qkv.weightQ4_0[7168 21504]
-
blk.53.attn_norm.biasF32[7168]
-
blk.53.attn_norm.weightF32[7168]
-
blk.53.ffn_down.biasF32[7168]
-
blk.53.ffn_down.weightQ4_0[28672 7168]
-
blk.53.ffn_up.biasF32[28672]
-
blk.53.ffn_up.weightQ4_0[7168 28672]
-
blk.53.ffn_norm.biasF32[7168]
-
blk.53.ffn_norm.weightF32[7168]
-
blk.53.attn_output.biasF32[7168]
-
blk.53.attn_output.weightQ4_0[7168 7168]
-
blk.53.attn_qkv.biasF32[21504]
-
blk.53.attn_qkv.weightQ4_0[7168 21504]
-
output.weightQ6_K[7168 128000]
-
output_norm.biasF32[7168]
-
output_norm.weightF32[7168]
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
blk.20
blk.21
blk.22
blk.23
blk.24
blk.25
blk.26
blk.27
blk.28
blk.29
blk.30
blk.31
blk.32
blk.33
blk.34
blk.35
blk.36
blk.37
blk.38
blk.39
blk.40
blk.41
blk.42
blk.43
blk.44
blk.45
blk.46
blk.47
blk.48
blk.49
blk.50
blk.51
blk.52
blk.53