-
general.architecture
chatglm
-
general.file_type
Q3_K_S
-
general.quantization_version
2
-
chatglm.attention.head_count
32
-
chatglm.attention.head_count_kv
2
-
chatglm.attention.layer_norm_rms_epsilon
1.5625e-07
-
chatglm.block_count
40
-
chatglm.context_length
8192
-
chatglm.embedding_length
4096
-
chatglm.feed_forward_length
13696
-
chatglm.rope.dimension_count
64
-
chatglm.rope.freq_base
10000
-
tokenizer.ggml.add_bos_token
false
-
tokenizer.ggml.eos_token_id
151329
-
tokenizer.ggml.eot_token_id
151336
-
tokenizer.ggml.merges
[Ġ Ġ ĠĠ ĠĠ i n Ġ t ĠĠĠĠ ĠĠĠĠ ...]
-
tokenizer.ggml.model
gpt2
-
tokenizer.ggml.padding_token_id
151329
-
tokenizer.ggml.pre
chatglm-bpe
-
tokenizer.ggml.token_type
[1 1 1 1 1 ...]
-
tokenizer.ggml.tokens
[! " # $ % ...]
-
tokenizer.ggml.unknown_token_id
151329
-
Name
Type
Shape
-
token_embd.weight
Q3_K
[4096 151552]
-
blk.0.attn_norm.weight
F32
[4096]
-
blk.0.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.0.ffn_up.weight
Q3_K
[4096 27392]
-
blk.0.ffn_norm.weight
F32
[4096]
-
blk.0.attn_output.weight
Q3_K
[4096 4096]
-
blk.0.attn_qkv.bias
F32
[4608]
-
blk.0.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.1.attn_norm.weight
F32
[4096]
-
blk.1.ffn_up.weight
Q3_K
[4096 27392]
-
blk.1.ffn_norm.weight
F32
[4096]
-
blk.1.attn_output.weight
Q3_K
[4096 4096]
-
blk.1.attn_qkv.bias
F32
[4608]
-
blk.1.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.1.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.2.attn_norm.weight
F32
[4096]
-
blk.2.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.2.ffn_up.weight
Q3_K
[4096 27392]
-
blk.2.ffn_norm.weight
F32
[4096]
-
blk.2.attn_output.weight
Q3_K
[4096 4096]
-
blk.2.attn_qkv.bias
F32
[4608]
-
blk.2.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.3.attn_norm.weight
F32
[4096]
-
blk.3.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.3.ffn_up.weight
Q3_K
[4096 27392]
-
blk.3.ffn_norm.weight
F32
[4096]
-
blk.3.attn_output.weight
Q3_K
[4096 4096]
-
blk.3.attn_qkv.bias
F32
[4608]
-
blk.3.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.4.attn_norm.weight
F32
[4096]
-
blk.4.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.4.ffn_up.weight
Q3_K
[4096 27392]
-
blk.4.ffn_norm.weight
F32
[4096]
-
blk.4.attn_output.weight
Q3_K
[4096 4096]
-
blk.4.attn_qkv.bias
F32
[4608]
-
blk.4.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.5.attn_norm.weight
F32
[4096]
-
blk.5.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.5.ffn_up.weight
Q3_K
[4096 27392]
-
blk.5.ffn_norm.weight
F32
[4096]
-
blk.5.attn_output.weight
Q3_K
[4096 4096]
-
blk.5.attn_qkv.bias
F32
[4608]
-
blk.5.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.6.attn_norm.weight
F32
[4096]
-
blk.6.ffn_norm.weight
F32
[4096]
-
blk.6.attn_output.weight
Q3_K
[4096 4096]
-
blk.6.attn_qkv.bias
F32
[4608]
-
blk.6.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.6.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.6.ffn_up.weight
Q3_K
[4096 27392]
-
blk.7.attn_norm.weight
F32
[4096]
-
blk.7.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.7.ffn_up.weight
Q3_K
[4096 27392]
-
blk.7.ffn_norm.weight
F32
[4096]
-
blk.7.attn_output.weight
Q3_K
[4096 4096]
-
blk.7.attn_qkv.bias
F32
[4608]
-
blk.7.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.8.attn_norm.weight
F32
[4096]
-
blk.8.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.8.ffn_up.weight
Q3_K
[4096 27392]
-
blk.8.ffn_norm.weight
F32
[4096]
-
blk.8.attn_output.weight
Q3_K
[4096 4096]
-
blk.8.attn_qkv.bias
F32
[4608]
-
blk.8.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.9.attn_norm.weight
F32
[4096]
-
blk.9.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.9.ffn_up.weight
Q3_K
[4096 27392]
-
blk.9.ffn_norm.weight
F32
[4096]
-
blk.9.attn_output.weight
Q3_K
[4096 4096]
-
blk.9.attn_qkv.bias
F32
[4608]
-
blk.9.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.10.attn_norm.weight
F32
[4096]
-
blk.10.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.10.ffn_up.weight
Q3_K
[4096 27392]
-
blk.10.ffn_norm.weight
F32
[4096]
-
blk.10.attn_output.weight
Q3_K
[4096 4096]
-
blk.10.attn_qkv.bias
F32
[4608]
-
blk.10.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.11.attn_norm.weight
F32
[4096]
-
blk.11.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.11.ffn_up.weight
Q3_K
[4096 27392]
-
blk.11.ffn_norm.weight
F32
[4096]
-
blk.11.attn_output.weight
Q3_K
[4096 4096]
-
blk.11.attn_qkv.bias
F32
[4608]
-
blk.11.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.12.attn_norm.weight
F32
[4096]
-
blk.12.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.12.ffn_up.weight
Q3_K
[4096 27392]
-
blk.12.ffn_norm.weight
F32
[4096]
-
blk.12.attn_output.weight
Q3_K
[4096 4096]
-
blk.12.attn_qkv.bias
F32
[4608]
-
blk.12.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.13.attn_norm.weight
F32
[4096]
-
blk.13.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.13.ffn_up.weight
Q3_K
[4096 27392]
-
blk.13.ffn_norm.weight
F32
[4096]
-
blk.13.attn_output.weight
Q3_K
[4096 4096]
-
blk.13.attn_qkv.bias
F32
[4608]
-
blk.13.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.14.attn_norm.weight
F32
[4096]
-
blk.14.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.14.ffn_up.weight
Q3_K
[4096 27392]
-
blk.14.ffn_norm.weight
F32
[4096]
-
blk.14.attn_output.weight
Q3_K
[4096 4096]
-
blk.14.attn_qkv.bias
F32
[4608]
-
blk.14.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.15.attn_norm.weight
F32
[4096]
-
blk.15.ffn_up.weight
Q3_K
[4096 27392]
-
blk.15.ffn_norm.weight
F32
[4096]
-
blk.15.attn_output.weight
Q3_K
[4096 4096]
-
blk.15.attn_qkv.bias
F32
[4608]
-
blk.15.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.15.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.16.attn_norm.weight
F32
[4096]
-
blk.16.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.16.ffn_up.weight
Q3_K
[4096 27392]
-
blk.16.ffn_norm.weight
F32
[4096]
-
blk.16.attn_output.weight
Q3_K
[4096 4096]
-
blk.16.attn_qkv.bias
F32
[4608]
-
blk.16.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.17.attn_norm.weight
F32
[4096]
-
blk.17.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.17.ffn_up.weight
Q3_K
[4096 27392]
-
blk.17.ffn_norm.weight
F32
[4096]
-
blk.17.attn_output.weight
Q3_K
[4096 4096]
-
blk.17.attn_qkv.bias
F32
[4608]
-
blk.17.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.18.attn_norm.weight
F32
[4096]
-
blk.18.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.18.ffn_up.weight
Q3_K
[4096 27392]
-
blk.18.ffn_norm.weight
F32
[4096]
-
blk.18.attn_output.weight
Q3_K
[4096 4096]
-
blk.18.attn_qkv.bias
F32
[4608]
-
blk.18.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.19.attn_norm.weight
F32
[4096]
-
blk.19.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.19.ffn_up.weight
Q3_K
[4096 27392]
-
blk.19.ffn_norm.weight
F32
[4096]
-
blk.19.attn_output.weight
Q3_K
[4096 4096]
-
blk.19.attn_qkv.bias
F32
[4608]
-
blk.19.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.20.attn_norm.weight
F32
[4096]
-
blk.20.ffn_norm.weight
F32
[4096]
-
blk.20.attn_output.weight
Q3_K
[4096 4096]
-
blk.20.attn_qkv.bias
F32
[4608]
-
blk.20.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.20.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.20.ffn_up.weight
Q3_K
[4096 27392]
-
blk.21.attn_norm.weight
F32
[4096]
-
blk.21.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.21.ffn_up.weight
Q3_K
[4096 27392]
-
blk.21.ffn_norm.weight
F32
[4096]
-
blk.21.attn_output.weight
Q3_K
[4096 4096]
-
blk.21.attn_qkv.bias
F32
[4608]
-
blk.21.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.22.attn_norm.weight
F32
[4096]
-
blk.22.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.22.ffn_up.weight
Q3_K
[4096 27392]
-
blk.22.ffn_norm.weight
F32
[4096]
-
blk.22.attn_output.weight
Q3_K
[4096 4096]
-
blk.22.attn_qkv.bias
F32
[4608]
-
blk.22.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.23.attn_norm.weight
F32
[4096]
-
blk.23.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.23.ffn_up.weight
Q3_K
[4096 27392]
-
blk.23.ffn_norm.weight
F32
[4096]
-
blk.23.attn_output.weight
Q3_K
[4096 4096]
-
blk.23.attn_qkv.bias
F32
[4608]
-
blk.23.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.24.attn_norm.weight
F32
[4096]
-
blk.24.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.24.ffn_up.weight
Q3_K
[4096 27392]
-
blk.24.ffn_norm.weight
F32
[4096]
-
blk.24.attn_output.weight
Q3_K
[4096 4096]
-
blk.24.attn_qkv.bias
F32
[4608]
-
blk.24.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.25.attn_norm.weight
F32
[4096]
-
blk.25.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.25.ffn_up.weight
Q3_K
[4096 27392]
-
blk.25.ffn_norm.weight
F32
[4096]
-
blk.25.attn_output.weight
Q3_K
[4096 4096]
-
blk.25.attn_qkv.bias
F32
[4608]
-
blk.25.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.26.attn_norm.weight
F32
[4096]
-
blk.26.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.26.ffn_up.weight
Q3_K
[4096 27392]
-
blk.26.ffn_norm.weight
F32
[4096]
-
blk.26.attn_output.weight
Q3_K
[4096 4096]
-
blk.26.attn_qkv.bias
F32
[4608]
-
blk.26.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.27.attn_norm.weight
F32
[4096]
-
blk.27.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.27.ffn_up.weight
Q3_K
[4096 27392]
-
blk.27.ffn_norm.weight
F32
[4096]
-
blk.27.attn_output.weight
Q3_K
[4096 4096]
-
blk.27.attn_qkv.bias
F32
[4608]
-
blk.27.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.28.attn_norm.weight
F32
[4096]
-
blk.28.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.28.ffn_up.weight
Q3_K
[4096 27392]
-
blk.28.ffn_norm.weight
F32
[4096]
-
blk.28.attn_output.weight
Q3_K
[4096 4096]
-
blk.28.attn_qkv.bias
F32
[4608]
-
blk.28.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.29.attn_norm.weight
F32
[4096]
-
blk.29.ffn_up.weight
Q3_K
[4096 27392]
-
blk.29.ffn_norm.weight
F32
[4096]
-
blk.29.attn_output.weight
Q3_K
[4096 4096]
-
blk.29.attn_qkv.bias
F32
[4608]
-
blk.29.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.29.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.30.attn_norm.weight
F32
[4096]
-
blk.30.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.30.ffn_up.weight
Q3_K
[4096 27392]
-
blk.30.ffn_norm.weight
F32
[4096]
-
blk.30.attn_output.weight
Q3_K
[4096 4096]
-
blk.30.attn_qkv.bias
F32
[4608]
-
blk.30.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.31.attn_norm.weight
F32
[4096]
-
blk.31.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.31.ffn_up.weight
Q3_K
[4096 27392]
-
blk.31.ffn_norm.weight
F32
[4096]
-
blk.31.attn_output.weight
Q3_K
[4096 4096]
-
blk.31.attn_qkv.bias
F32
[4608]
-
blk.31.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.32.attn_norm.weight
F32
[4096]
-
blk.32.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.32.ffn_up.weight
Q3_K
[4096 27392]
-
blk.32.ffn_norm.weight
F32
[4096]
-
blk.32.attn_output.weight
Q3_K
[4096 4096]
-
blk.32.attn_qkv.bias
F32
[4608]
-
blk.32.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.33.attn_norm.weight
F32
[4096]
-
blk.33.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.33.ffn_up.weight
Q3_K
[4096 27392]
-
blk.33.ffn_norm.weight
F32
[4096]
-
blk.33.attn_output.weight
Q3_K
[4096 4096]
-
blk.33.attn_qkv.bias
F32
[4608]
-
blk.33.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.34.attn_norm.weight
F32
[4096]
-
blk.34.ffn_norm.weight
F32
[4096]
-
blk.34.attn_output.weight
Q3_K
[4096 4096]
-
blk.34.attn_qkv.bias
F32
[4608]
-
blk.34.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.34.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.34.ffn_up.weight
Q3_K
[4096 27392]
-
blk.35.attn_norm.weight
F32
[4096]
-
blk.35.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.35.ffn_up.weight
Q3_K
[4096 27392]
-
blk.35.ffn_norm.weight
F32
[4096]
-
blk.35.attn_output.weight
Q3_K
[4096 4096]
-
blk.35.attn_qkv.bias
F32
[4608]
-
blk.35.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.36.attn_norm.weight
F32
[4096]
-
blk.36.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.36.ffn_up.weight
Q3_K
[4096 27392]
-
blk.36.ffn_norm.weight
F32
[4096]
-
blk.36.attn_output.weight
Q3_K
[4096 4096]
-
blk.36.attn_qkv.bias
F32
[4608]
-
blk.36.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.37.attn_norm.weight
F32
[4096]
-
blk.37.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.37.ffn_up.weight
Q3_K
[4096 27392]
-
blk.37.ffn_norm.weight
F32
[4096]
-
blk.37.attn_output.weight
Q3_K
[4096 4096]
-
blk.37.attn_qkv.bias
F32
[4608]
-
blk.37.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.38.attn_norm.weight
F32
[4096]
-
blk.38.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.38.ffn_up.weight
Q3_K
[4096 27392]
-
blk.38.ffn_norm.weight
F32
[4096]
-
blk.38.attn_output.weight
Q3_K
[4096 4096]
-
blk.38.attn_qkv.bias
F32
[4608]
-
blk.38.attn_qkv.weight
Q3_K
[4096 4608]
-
blk.39.attn_norm.weight
F32
[4096]
-
blk.39.ffn_down.weight
(!unknown_type 20!)
[13696 4096]
-
blk.39.ffn_up.weight
Q3_K
[4096 27392]
-
blk.39.ffn_norm.weight
F32
[4096]
-
blk.39.attn_output.weight
Q3_K
[4096 4096]
-
blk.39.attn_qkv.bias
F32
[4608]
-
blk.39.attn_qkv.weight
Q3_K
[4096 4608]
-
output_norm.weight
F32
[4096]
-
output.weight
Q6_K
[4096 151552]