-
general.architecture
llama
-
general.file_type
Q2_K
-
llama.attention.head_count
32
-
llama.attention.head_count_kv
8
-
llama.attention.key_length
128
-
llama.attention.layer_norm_rms_epsilon
1e-05
-
llama.attention.value_length
128
-
llama.block_count
40
-
llama.context_length
1024000
-
llama.embedding_length
5120
-
llama.feed_forward_length
14336
-
llama.rope.dimension_count
128
-
llama.rope.freq_base
1e+06
-
llama.vocab_size
131072
-
tokenizer.ggml.add_bos_token
true
-
tokenizer.ggml.add_eos_token
false
-
tokenizer.ggml.add_space_prefix
false
-
tokenizer.ggml.bos_token_id
1
-
tokenizer.ggml.eos_token_id
2
-
tokenizer.ggml.merges
[Ġ Ġ Ġ t e r i n Ġ ĠĠĠ ...]
-
tokenizer.ggml.model
gpt2
-
tokenizer.ggml.pre
tekken
-
tokenizer.ggml.token_type
[3 3 3 3 3 ...]
-
tokenizer.ggml.tokens
[<unk> <s> </s> [INST] [/INST] ...]
-
tokenizer.ggml.unknown_token_id
0
-
Name
Type
Shape
-
token_embd.weight
Q2_K
[5120 131072]
-
blk.0.attn_norm.weight
F32
[5120]
-
blk.0.ffn_down.weight
Q3_K
[14336 5120]
-
blk.0.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.0.ffn_up.weight
Q2_K
[5120 14336]
-
blk.0.ffn_norm.weight
F32
[5120]
-
blk.0.attn_k.weight
Q2_K
[5120 1024]
-
blk.0.attn_output.weight
Q3_K
[4096 5120]
-
blk.0.attn_q.weight
Q2_K
[5120 4096]
-
blk.0.attn_v.weight
Q4_K
[5120 1024]
-
blk.1.attn_norm.weight
F32
[5120]
-
blk.1.ffn_down.weight
Q3_K
[14336 5120]
-
blk.1.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.1.ffn_up.weight
Q2_K
[5120 14336]
-
blk.1.ffn_norm.weight
F32
[5120]
-
blk.1.attn_k.weight
Q2_K
[5120 1024]
-
blk.1.attn_output.weight
Q3_K
[4096 5120]
-
blk.1.attn_q.weight
Q2_K
[5120 4096]
-
blk.1.attn_v.weight
Q4_K
[5120 1024]
-
blk.2.attn_norm.weight
F32
[5120]
-
blk.2.ffn_down.weight
Q3_K
[14336 5120]
-
blk.2.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.2.ffn_up.weight
Q2_K
[5120 14336]
-
blk.2.ffn_norm.weight
F32
[5120]
-
blk.2.attn_k.weight
Q2_K
[5120 1024]
-
blk.2.attn_output.weight
Q3_K
[4096 5120]
-
blk.2.attn_q.weight
Q2_K
[5120 4096]
-
blk.2.attn_v.weight
Q4_K
[5120 1024]
-
blk.3.attn_norm.weight
F32
[5120]
-
blk.3.ffn_down.weight
Q3_K
[14336 5120]
-
blk.3.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.3.ffn_up.weight
Q2_K
[5120 14336]
-
blk.3.ffn_norm.weight
F32
[5120]
-
blk.3.attn_k.weight
Q2_K
[5120 1024]
-
blk.3.attn_output.weight
Q3_K
[4096 5120]
-
blk.3.attn_q.weight
Q2_K
[5120 4096]
-
blk.3.attn_v.weight
Q4_K
[5120 1024]
-
blk.4.attn_norm.weight
F32
[5120]
-
blk.4.ffn_down.weight
Q3_K
[14336 5120]
-
blk.4.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.4.ffn_up.weight
Q2_K
[5120 14336]
-
blk.4.ffn_norm.weight
F32
[5120]
-
blk.4.attn_k.weight
Q2_K
[5120 1024]
-
blk.4.attn_output.weight
Q3_K
[4096 5120]
-
blk.4.attn_q.weight
Q2_K
[5120 4096]
-
blk.4.attn_v.weight
Q4_K
[5120 1024]
-
blk.5.attn_norm.weight
F32
[5120]
-
blk.5.ffn_down.weight
Q3_K
[14336 5120]
-
blk.5.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.5.ffn_up.weight
Q2_K
[5120 14336]
-
blk.5.ffn_norm.weight
F32
[5120]
-
blk.5.attn_k.weight
Q2_K
[5120 1024]
-
blk.5.attn_output.weight
Q3_K
[4096 5120]
-
blk.5.attn_q.weight
Q2_K
[5120 4096]
-
blk.5.attn_v.weight
Q4_K
[5120 1024]
-
blk.6.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.6.attn_k.weight
Q2_K
[5120 1024]
-
blk.6.attn_output.weight
Q3_K
[4096 5120]
-
blk.6.attn_q.weight
Q2_K
[5120 4096]
-
blk.6.attn_v.weight
Q4_K
[5120 1024]
-
blk.6.attn_norm.weight
F32
[5120]
-
blk.6.ffn_down.weight
Q3_K
[14336 5120]
-
blk.6.ffn_up.weight
Q2_K
[5120 14336]
-
blk.6.ffn_norm.weight
F32
[5120]
-
blk.7.attn_norm.weight
F32
[5120]
-
blk.7.ffn_down.weight
Q3_K
[14336 5120]
-
blk.7.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.7.ffn_up.weight
Q2_K
[5120 14336]
-
blk.7.ffn_norm.weight
F32
[5120]
-
blk.7.attn_k.weight
Q2_K
[5120 1024]
-
blk.7.attn_output.weight
Q3_K
[4096 5120]
-
blk.7.attn_q.weight
Q2_K
[5120 4096]
-
blk.7.attn_v.weight
Q4_K
[5120 1024]
-
blk.8.attn_norm.weight
F32
[5120]
-
blk.8.ffn_down.weight
Q3_K
[14336 5120]
-
blk.8.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.8.ffn_up.weight
Q2_K
[5120 14336]
-
blk.8.ffn_norm.weight
F32
[5120]
-
blk.8.attn_k.weight
Q2_K
[5120 1024]
-
blk.8.attn_output.weight
Q3_K
[4096 5120]
-
blk.8.attn_q.weight
Q2_K
[5120 4096]
-
blk.8.attn_v.weight
Q4_K
[5120 1024]
-
blk.9.attn_norm.weight
F32
[5120]
-
blk.9.ffn_down.weight
Q3_K
[14336 5120]
-
blk.9.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.9.ffn_up.weight
Q2_K
[5120 14336]
-
blk.9.ffn_norm.weight
F32
[5120]
-
blk.9.attn_k.weight
Q2_K
[5120 1024]
-
blk.9.attn_output.weight
Q3_K
[4096 5120]
-
blk.9.attn_q.weight
Q2_K
[5120 4096]
-
blk.9.attn_v.weight
Q4_K
[5120 1024]
-
blk.10.attn_norm.weight
F32
[5120]
-
blk.10.ffn_down.weight
Q3_K
[14336 5120]
-
blk.10.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.10.ffn_up.weight
Q2_K
[5120 14336]
-
blk.10.ffn_norm.weight
F32
[5120]
-
blk.10.attn_k.weight
Q2_K
[5120 1024]
-
blk.10.attn_output.weight
Q3_K
[4096 5120]
-
blk.10.attn_q.weight
Q2_K
[5120 4096]
-
blk.10.attn_v.weight
Q4_K
[5120 1024]
-
blk.11.attn_norm.weight
F32
[5120]
-
blk.11.ffn_down.weight
Q3_K
[14336 5120]
-
blk.11.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.11.ffn_up.weight
Q2_K
[5120 14336]
-
blk.11.ffn_norm.weight
F32
[5120]
-
blk.11.attn_k.weight
Q2_K
[5120 1024]
-
blk.11.attn_output.weight
Q3_K
[4096 5120]
-
blk.11.attn_q.weight
Q2_K
[5120 4096]
-
blk.11.attn_v.weight
Q4_K
[5120 1024]
-
blk.12.attn_norm.weight
F32
[5120]
-
blk.12.ffn_down.weight
Q3_K
[14336 5120]
-
blk.12.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.12.ffn_up.weight
Q2_K
[5120 14336]
-
blk.12.ffn_norm.weight
F32
[5120]
-
blk.12.attn_k.weight
Q2_K
[5120 1024]
-
blk.12.attn_output.weight
Q3_K
[4096 5120]
-
blk.12.attn_q.weight
Q2_K
[5120 4096]
-
blk.12.attn_v.weight
Q4_K
[5120 1024]
-
blk.13.attn_norm.weight
F32
[5120]
-
blk.13.ffn_down.weight
Q3_K
[14336 5120]
-
blk.13.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.13.ffn_up.weight
Q2_K
[5120 14336]
-
blk.13.ffn_norm.weight
F32
[5120]
-
blk.13.attn_k.weight
Q2_K
[5120 1024]
-
blk.13.attn_output.weight
Q3_K
[4096 5120]
-
blk.13.attn_q.weight
Q2_K
[5120 4096]
-
blk.13.attn_v.weight
Q4_K
[5120 1024]
-
blk.14.attn_norm.weight
F32
[5120]
-
blk.14.ffn_down.weight
Q3_K
[14336 5120]
-
blk.14.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.14.ffn_up.weight
Q2_K
[5120 14336]
-
blk.14.ffn_norm.weight
F32
[5120]
-
blk.14.attn_k.weight
Q2_K
[5120 1024]
-
blk.14.attn_output.weight
Q3_K
[4096 5120]
-
blk.14.attn_q.weight
Q2_K
[5120 4096]
-
blk.14.attn_v.weight
Q4_K
[5120 1024]
-
blk.15.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.15.attn_k.weight
Q2_K
[5120 1024]
-
blk.15.attn_output.weight
Q3_K
[4096 5120]
-
blk.15.attn_q.weight
Q2_K
[5120 4096]
-
blk.15.attn_v.weight
Q4_K
[5120 1024]
-
blk.15.attn_norm.weight
F32
[5120]
-
blk.15.ffn_down.weight
Q3_K
[14336 5120]
-
blk.15.ffn_up.weight
Q2_K
[5120 14336]
-
blk.15.ffn_norm.weight
F32
[5120]
-
blk.16.attn_norm.weight
F32
[5120]
-
blk.16.ffn_down.weight
Q3_K
[14336 5120]
-
blk.16.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.16.ffn_up.weight
Q2_K
[5120 14336]
-
blk.16.ffn_norm.weight
F32
[5120]
-
blk.16.attn_k.weight
Q2_K
[5120 1024]
-
blk.16.attn_output.weight
Q3_K
[4096 5120]
-
blk.16.attn_q.weight
Q2_K
[5120 4096]
-
blk.16.attn_v.weight
Q4_K
[5120 1024]
-
blk.17.attn_norm.weight
F32
[5120]
-
blk.17.ffn_down.weight
Q3_K
[14336 5120]
-
blk.17.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.17.ffn_up.weight
Q2_K
[5120 14336]
-
blk.17.ffn_norm.weight
F32
[5120]
-
blk.17.attn_k.weight
Q2_K
[5120 1024]
-
blk.17.attn_output.weight
Q3_K
[4096 5120]
-
blk.17.attn_q.weight
Q2_K
[5120 4096]
-
blk.17.attn_v.weight
Q4_K
[5120 1024]
-
blk.18.attn_norm.weight
F32
[5120]
-
blk.18.ffn_down.weight
Q3_K
[14336 5120]
-
blk.18.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.18.ffn_up.weight
Q2_K
[5120 14336]
-
blk.18.ffn_norm.weight
F32
[5120]
-
blk.18.attn_k.weight
Q2_K
[5120 1024]
-
blk.18.attn_output.weight
Q3_K
[4096 5120]
-
blk.18.attn_q.weight
Q2_K
[5120 4096]
-
blk.18.attn_v.weight
Q4_K
[5120 1024]
-
blk.19.attn_norm.weight
F32
[5120]
-
blk.19.ffn_down.weight
Q3_K
[14336 5120]
-
blk.19.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.19.ffn_up.weight
Q2_K
[5120 14336]
-
blk.19.ffn_norm.weight
F32
[5120]
-
blk.19.attn_k.weight
Q2_K
[5120 1024]
-
blk.19.attn_output.weight
Q3_K
[4096 5120]
-
blk.19.attn_q.weight
Q2_K
[5120 4096]
-
blk.19.attn_v.weight
Q4_K
[5120 1024]
-
blk.20.attn_norm.weight
F32
[5120]
-
blk.20.ffn_down.weight
Q3_K
[14336 5120]
-
blk.20.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.20.ffn_up.weight
Q2_K
[5120 14336]
-
blk.20.ffn_norm.weight
F32
[5120]
-
blk.20.attn_k.weight
Q2_K
[5120 1024]
-
blk.20.attn_output.weight
Q3_K
[4096 5120]
-
blk.20.attn_q.weight
Q2_K
[5120 4096]
-
blk.20.attn_v.weight
Q4_K
[5120 1024]
-
blk.21.attn_norm.weight
F32
[5120]
-
blk.21.ffn_down.weight
Q3_K
[14336 5120]
-
blk.21.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.21.ffn_up.weight
Q2_K
[5120 14336]
-
blk.21.ffn_norm.weight
F32
[5120]
-
blk.21.attn_k.weight
Q2_K
[5120 1024]
-
blk.21.attn_output.weight
Q3_K
[4096 5120]
-
blk.21.attn_q.weight
Q2_K
[5120 4096]
-
blk.21.attn_v.weight
Q4_K
[5120 1024]
-
blk.22.attn_norm.weight
F32
[5120]
-
blk.22.ffn_down.weight
Q3_K
[14336 5120]
-
blk.22.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.22.ffn_up.weight
Q2_K
[5120 14336]
-
blk.22.ffn_norm.weight
F32
[5120]
-
blk.22.attn_k.weight
Q2_K
[5120 1024]
-
blk.22.attn_output.weight
Q3_K
[4096 5120]
-
blk.22.attn_q.weight
Q2_K
[5120 4096]
-
blk.22.attn_v.weight
Q4_K
[5120 1024]
-
blk.23.attn_norm.weight
F32
[5120]
-
blk.23.ffn_down.weight
Q3_K
[14336 5120]
-
blk.23.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.23.ffn_up.weight
Q2_K
[5120 14336]
-
blk.23.ffn_norm.weight
F32
[5120]
-
blk.23.attn_k.weight
Q2_K
[5120 1024]
-
blk.23.attn_output.weight
Q3_K
[4096 5120]
-
blk.23.attn_q.weight
Q2_K
[5120 4096]
-
blk.23.attn_v.weight
Q4_K
[5120 1024]
-
blk.24.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.24.attn_k.weight
Q2_K
[5120 1024]
-
blk.24.attn_output.weight
Q3_K
[4096 5120]
-
blk.24.attn_q.weight
Q2_K
[5120 4096]
-
blk.24.attn_v.weight
Q4_K
[5120 1024]
-
blk.24.attn_norm.weight
F32
[5120]
-
blk.24.ffn_down.weight
Q3_K
[14336 5120]
-
blk.24.ffn_up.weight
Q2_K
[5120 14336]
-
blk.24.ffn_norm.weight
F32
[5120]
-
blk.25.attn_norm.weight
F32
[5120]
-
blk.25.ffn_down.weight
Q3_K
[14336 5120]
-
blk.25.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.25.ffn_up.weight
Q2_K
[5120 14336]
-
blk.25.ffn_norm.weight
F32
[5120]
-
blk.25.attn_k.weight
Q2_K
[5120 1024]
-
blk.25.attn_output.weight
Q3_K
[4096 5120]
-
blk.25.attn_q.weight
Q2_K
[5120 4096]
-
blk.25.attn_v.weight
Q4_K
[5120 1024]
-
blk.26.attn_norm.weight
F32
[5120]
-
blk.26.ffn_down.weight
Q3_K
[14336 5120]
-
blk.26.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.26.ffn_up.weight
Q2_K
[5120 14336]
-
blk.26.ffn_norm.weight
F32
[5120]
-
blk.26.attn_k.weight
Q2_K
[5120 1024]
-
blk.26.attn_output.weight
Q3_K
[4096 5120]
-
blk.26.attn_q.weight
Q2_K
[5120 4096]
-
blk.26.attn_v.weight
Q4_K
[5120 1024]
-
blk.27.attn_norm.weight
F32
[5120]
-
blk.27.ffn_down.weight
Q3_K
[14336 5120]
-
blk.27.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.27.ffn_up.weight
Q2_K
[5120 14336]
-
blk.27.ffn_norm.weight
F32
[5120]
-
blk.27.attn_k.weight
Q2_K
[5120 1024]
-
blk.27.attn_output.weight
Q3_K
[4096 5120]
-
blk.27.attn_q.weight
Q2_K
[5120 4096]
-
blk.27.attn_v.weight
Q4_K
[5120 1024]
-
blk.28.attn_norm.weight
F32
[5120]
-
blk.28.ffn_down.weight
Q3_K
[14336 5120]
-
blk.28.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.28.ffn_up.weight
Q2_K
[5120 14336]
-
blk.28.ffn_norm.weight
F32
[5120]
-
blk.28.attn_k.weight
Q2_K
[5120 1024]
-
blk.28.attn_output.weight
Q3_K
[4096 5120]
-
blk.28.attn_q.weight
Q2_K
[5120 4096]
-
blk.28.attn_v.weight
Q4_K
[5120 1024]
-
blk.29.attn_norm.weight
F32
[5120]
-
blk.29.ffn_down.weight
Q3_K
[14336 5120]
-
blk.29.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.29.ffn_up.weight
Q2_K
[5120 14336]
-
blk.29.ffn_norm.weight
F32
[5120]
-
blk.29.attn_k.weight
Q2_K
[5120 1024]
-
blk.29.attn_output.weight
Q3_K
[4096 5120]
-
blk.29.attn_q.weight
Q2_K
[5120 4096]
-
blk.29.attn_v.weight
Q4_K
[5120 1024]
-
blk.30.attn_norm.weight
F32
[5120]
-
blk.30.ffn_down.weight
Q3_K
[14336 5120]
-
blk.30.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.30.ffn_up.weight
Q2_K
[5120 14336]
-
blk.30.ffn_norm.weight
F32
[5120]
-
blk.30.attn_k.weight
Q2_K
[5120 1024]
-
blk.30.attn_output.weight
Q3_K
[4096 5120]
-
blk.30.attn_q.weight
Q2_K
[5120 4096]
-
blk.30.attn_v.weight
Q4_K
[5120 1024]
-
blk.31.attn_norm.weight
F32
[5120]
-
blk.31.ffn_down.weight
Q3_K
[14336 5120]
-
blk.31.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.31.ffn_up.weight
Q2_K
[5120 14336]
-
blk.31.ffn_norm.weight
F32
[5120]
-
blk.31.attn_k.weight
Q2_K
[5120 1024]
-
blk.31.attn_output.weight
Q3_K
[4096 5120]
-
blk.31.attn_q.weight
Q2_K
[5120 4096]
-
blk.31.attn_v.weight
Q4_K
[5120 1024]
-
blk.32.attn_norm.weight
F32
[5120]
-
blk.32.ffn_down.weight
Q3_K
[14336 5120]
-
blk.32.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.32.ffn_up.weight
Q2_K
[5120 14336]
-
blk.32.ffn_norm.weight
F32
[5120]
-
blk.32.attn_k.weight
Q2_K
[5120 1024]
-
blk.32.attn_output.weight
Q3_K
[4096 5120]
-
blk.32.attn_q.weight
Q2_K
[5120 4096]
-
blk.32.attn_v.weight
Q4_K
[5120 1024]
-
blk.33.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.33.attn_k.weight
Q2_K
[5120 1024]
-
blk.33.attn_output.weight
Q3_K
[4096 5120]
-
blk.33.attn_q.weight
Q2_K
[5120 4096]
-
blk.33.attn_v.weight
Q4_K
[5120 1024]
-
blk.33.attn_norm.weight
F32
[5120]
-
blk.33.ffn_down.weight
Q3_K
[14336 5120]
-
blk.33.ffn_up.weight
Q2_K
[5120 14336]
-
blk.33.ffn_norm.weight
F32
[5120]
-
blk.34.attn_norm.weight
F32
[5120]
-
blk.34.ffn_down.weight
Q3_K
[14336 5120]
-
blk.34.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.34.ffn_up.weight
Q2_K
[5120 14336]
-
blk.34.ffn_norm.weight
F32
[5120]
-
blk.34.attn_k.weight
Q2_K
[5120 1024]
-
blk.34.attn_output.weight
Q3_K
[4096 5120]
-
blk.34.attn_q.weight
Q2_K
[5120 4096]
-
blk.34.attn_v.weight
Q4_K
[5120 1024]
-
blk.35.attn_norm.weight
F32
[5120]
-
blk.35.ffn_down.weight
Q3_K
[14336 5120]
-
blk.35.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.35.ffn_up.weight
Q2_K
[5120 14336]
-
blk.35.ffn_norm.weight
F32
[5120]
-
blk.35.attn_k.weight
Q2_K
[5120 1024]
-
blk.35.attn_output.weight
Q3_K
[4096 5120]
-
blk.35.attn_q.weight
Q2_K
[5120 4096]
-
blk.35.attn_v.weight
Q4_K
[5120 1024]
-
blk.36.attn_norm.weight
F32
[5120]
-
blk.36.ffn_down.weight
Q3_K
[14336 5120]
-
blk.36.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.36.ffn_up.weight
Q2_K
[5120 14336]
-
blk.36.ffn_norm.weight
F32
[5120]
-
blk.36.attn_k.weight
Q2_K
[5120 1024]
-
blk.36.attn_output.weight
Q3_K
[4096 5120]
-
blk.36.attn_q.weight
Q2_K
[5120 4096]
-
blk.36.attn_v.weight
Q4_K
[5120 1024]
-
blk.37.attn_norm.weight
F32
[5120]
-
blk.37.ffn_down.weight
Q3_K
[14336 5120]
-
blk.37.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.37.ffn_up.weight
Q2_K
[5120 14336]
-
blk.37.ffn_norm.weight
F32
[5120]
-
blk.37.attn_k.weight
Q2_K
[5120 1024]
-
blk.37.attn_output.weight
Q3_K
[4096 5120]
-
blk.37.attn_q.weight
Q2_K
[5120 4096]
-
blk.37.attn_v.weight
Q4_K
[5120 1024]
-
blk.38.attn_norm.weight
F32
[5120]
-
blk.38.ffn_down.weight
Q3_K
[14336 5120]
-
blk.38.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.38.ffn_up.weight
Q2_K
[5120 14336]
-
blk.38.ffn_norm.weight
F32
[5120]
-
blk.38.attn_k.weight
Q2_K
[5120 1024]
-
blk.38.attn_output.weight
Q3_K
[4096 5120]
-
blk.38.attn_q.weight
Q2_K
[5120 4096]
-
blk.38.attn_v.weight
Q4_K
[5120 1024]
-
blk.39.attn_norm.weight
F32
[5120]
-
blk.39.ffn_down.weight
Q3_K
[14336 5120]
-
blk.39.ffn_gate.weight
Q2_K
[5120 14336]
-
blk.39.ffn_up.weight
Q2_K
[5120 14336]
-
blk.39.ffn_norm.weight
F32
[5120]
-
blk.39.attn_k.weight
Q2_K
[5120 1024]
-
blk.39.attn_output.weight
Q3_K
[4096 5120]
-
blk.39.attn_q.weight
Q2_K
[5120 4096]
-
blk.39.attn_v.weight
Q4_K
[5120 1024]
-
output.weight
Q6_K
[5120 131072]
-
output_norm.weight
F32
[5120]