Models
Docs
Pricing
Sign in
Download
Models
Download
Docs
Pricing
Sign in
qwen3-coder
:30b-a3b-q8_0
8.3M
Downloads
Updated
10 months ago
Alibaba's performant long context models for agentic and coding tasks.
Alibaba's performant long context models for agentic and coding tasks.
Cancel
tools
30b
480b
qwen3-coder:30b-a3b-q8_0
...
/
model
1a8a72d99ed2 · 32GB
Metadata
general.architecture
qwen3moe
qwen3moe
general.file_type
Q8_0
Q8_0
qwen3moe.attention.head_count
32
32
qwen3moe.attention.head_count_kv
4
4
qwen3moe.attention.key_length
128
128
qwen3moe.attention.layer_norm_rms_epsilon
1e-06
1e-06
qwen3moe.attention.value_length
128
128
qwen3moe.block_count
48
48
qwen3moe.context_length
262144
262144
qwen3moe.embedding_length
2048
2048
qwen3moe.expert_count
128
128
qwen3moe.expert_feed_forward_length
768
768
qwen3moe.expert_shared_feed_forward_length
0
0
qwen3moe.expert_used_count
8
8
qwen3moe.feed_forward_length
5472
5472
qwen3moe.rope.freq_base
1e+07
1e+07
tokenizer.ggml.add_bos_token
false
false
tokenizer.ggml.eos_token_id
151645
151645
tokenizer.ggml.merges
[Ġ Ġ, ĠĠ ĠĠ, i n, Ġ t, ĠĠĠĠ ĠĠĠĠ, ...]
[Ġ Ġ, ĠĠ ĠĠ, i n, Ġ t, ĠĠĠĠ ĠĠĠĠ, ...]
tokenizer.ggml.model
gpt2
gpt2
tokenizer.ggml.padding_token_id
151643
151643
tokenizer.ggml.pre
qwen2
qwen2
tokenizer.ggml.token_type
[1, 1, 1, 1, 1, ...]
[1, 1, 1, 1, 1, ...]
tokenizer.ggml.tokens
[!, ", #, $, %, ...]
[!, ", #, $, %, ...]
Tensor
Name
Type
Shape
token_embd.weight
Q8_0
Q8_0
[2048, 151936]
blk.0
blk.0.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.0.attn_k_norm.weight
F32
F32
[128]
blk.0.attn_norm.weight
F32
F32
[2048]
blk.0.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.0.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.0.attn_q_norm.weight
F32
F32
[128]
blk.0.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.0.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.0.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.0.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.0.ffn_norm.weight
F32
F32
[2048]
blk.0.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.1
blk.1.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.1.attn_k_norm.weight
F32
F32
[128]
blk.1.attn_norm.weight
F32
F32
[2048]
blk.1.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.1.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.1.attn_q_norm.weight
F32
F32
[128]
blk.1.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.1.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.1.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.1.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.1.ffn_norm.weight
F32
F32
[2048]
blk.1.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.2
blk.2.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.2.attn_k_norm.weight
F32
F32
[128]
blk.2.attn_norm.weight
F32
F32
[2048]
blk.2.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.2.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.2.attn_q_norm.weight
F32
F32
[128]
blk.2.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.2.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.2.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.2.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.2.ffn_norm.weight
F32
F32
[2048]
blk.2.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.3
blk.3.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.3.attn_k_norm.weight
F32
F32
[128]
blk.3.attn_norm.weight
F32
F32
[2048]
blk.3.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.3.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.3.attn_q_norm.weight
F32
F32
[128]
blk.3.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.3.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.3.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.3.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.3.ffn_norm.weight
F32
F32
[2048]
blk.3.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.4
blk.4.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.4.attn_k_norm.weight
F32
F32
[128]
blk.4.attn_norm.weight
F32
F32
[2048]
blk.4.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.4.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.4.attn_q_norm.weight
F32
F32
[128]
blk.4.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.4.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.4.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.4.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.4.ffn_norm.weight
F32
F32
[2048]
blk.4.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.5
blk.5.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.5.attn_k_norm.weight
F32
F32
[128]
blk.5.attn_norm.weight
F32
F32
[2048]
blk.5.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.5.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.5.attn_q_norm.weight
F32
F32
[128]
blk.5.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.5.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.5.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.5.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.5.ffn_norm.weight
F32
F32
[2048]
blk.5.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.6
blk.6.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.6.attn_k_norm.weight
F32
F32
[128]
blk.6.attn_norm.weight
F32
F32
[2048]
blk.6.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.6.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.6.attn_q_norm.weight
F32
F32
[128]
blk.6.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.6.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.6.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.6.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.6.ffn_norm.weight
F32
F32
[2048]
blk.6.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.7
blk.7.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.7.attn_k_norm.weight
F32
F32
[128]
blk.7.attn_norm.weight
F32
F32
[2048]
blk.7.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.7.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.7.attn_q_norm.weight
F32
F32
[128]
blk.7.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.7.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.7.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.7.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.7.ffn_norm.weight
F32
F32
[2048]
blk.7.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.8
blk.8.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.8.attn_k_norm.weight
F32
F32
[128]
blk.8.attn_norm.weight
F32
F32
[2048]
blk.8.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.8.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.8.attn_q_norm.weight
F32
F32
[128]
blk.8.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.8.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.8.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.8.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.8.ffn_norm.weight
F32
F32
[2048]
blk.8.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.9
blk.9.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.9.attn_k_norm.weight
F32
F32
[128]
blk.9.attn_norm.weight
F32
F32
[2048]
blk.9.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.9.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.9.attn_q_norm.weight
F32
F32
[128]
blk.9.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.9.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.9.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.9.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.9.ffn_norm.weight
F32
F32
[2048]
blk.9.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.10
blk.10.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.10.attn_k_norm.weight
F32
F32
[128]
blk.10.attn_norm.weight
F32
F32
[2048]
blk.10.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.10.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.10.attn_q_norm.weight
F32
F32
[128]
blk.10.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.10.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.10.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.10.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.10.ffn_norm.weight
F32
F32
[2048]
blk.10.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.11
blk.11.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.11.attn_k_norm.weight
F32
F32
[128]
blk.11.attn_norm.weight
F32
F32
[2048]
blk.11.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.11.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.11.attn_q_norm.weight
F32
F32
[128]
blk.11.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.11.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.11.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.11.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.11.ffn_norm.weight
F32
F32
[2048]
blk.11.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.12
blk.12.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.12.attn_k_norm.weight
F32
F32
[128]
blk.12.attn_norm.weight
F32
F32
[2048]
blk.12.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.12.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.12.attn_q_norm.weight
F32
F32
[128]
blk.12.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.12.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.12.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.12.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.12.ffn_norm.weight
F32
F32
[2048]
blk.12.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.13
blk.13.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.13.attn_k_norm.weight
F32
F32
[128]
blk.13.attn_norm.weight
F32
F32
[2048]
blk.13.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.13.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.13.attn_q_norm.weight
F32
F32
[128]
blk.13.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.13.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.13.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.13.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.13.ffn_norm.weight
F32
F32
[2048]
blk.13.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.14
blk.14.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.14.attn_k_norm.weight
F32
F32
[128]
blk.14.attn_norm.weight
F32
F32
[2048]
blk.14.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.14.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.14.attn_q_norm.weight
F32
F32
[128]
blk.14.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.14.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.14.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.14.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.14.ffn_norm.weight
F32
F32
[2048]
blk.14.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.15
blk.15.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.15.attn_k_norm.weight
F32
F32
[128]
blk.15.attn_norm.weight
F32
F32
[2048]
blk.15.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.15.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.15.attn_q_norm.weight
F32
F32
[128]
blk.15.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.15.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.15.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.15.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.15.ffn_norm.weight
F32
F32
[2048]
blk.15.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.16
blk.16.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.16.attn_k_norm.weight
F32
F32
[128]
blk.16.attn_norm.weight
F32
F32
[2048]
blk.16.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.16.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.16.attn_q_norm.weight
F32
F32
[128]
blk.16.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.16.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.16.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.16.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.16.ffn_norm.weight
F32
F32
[2048]
blk.16.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.17
blk.17.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.17.attn_k_norm.weight
F32
F32
[128]
blk.17.attn_norm.weight
F32
F32
[2048]
blk.17.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.17.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.17.attn_q_norm.weight
F32
F32
[128]
blk.17.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.17.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.17.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.17.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.17.ffn_norm.weight
F32
F32
[2048]
blk.17.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.18
blk.18.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.18.attn_k_norm.weight
F32
F32
[128]
blk.18.attn_norm.weight
F32
F32
[2048]
blk.18.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.18.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.18.attn_q_norm.weight
F32
F32
[128]
blk.18.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.18.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.18.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.18.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.18.ffn_norm.weight
F32
F32
[2048]
blk.18.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.19
blk.19.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.19.attn_k_norm.weight
F32
F32
[128]
blk.19.attn_norm.weight
F32
F32
[2048]
blk.19.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.19.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.19.attn_q_norm.weight
F32
F32
[128]
blk.19.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.19.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.19.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.19.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.19.ffn_norm.weight
F32
F32
[2048]
blk.19.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.20
blk.20.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.20.attn_k_norm.weight
F32
F32
[128]
blk.20.attn_norm.weight
F32
F32
[2048]
blk.20.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.20.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.20.attn_q_norm.weight
F32
F32
[128]
blk.20.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.20.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.20.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.20.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.20.ffn_norm.weight
F32
F32
[2048]
blk.20.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.21
blk.21.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.21.attn_k_norm.weight
F32
F32
[128]
blk.21.attn_norm.weight
F32
F32
[2048]
blk.21.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.21.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.21.attn_q_norm.weight
F32
F32
[128]
blk.21.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.21.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.21.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.21.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.21.ffn_norm.weight
F32
F32
[2048]
blk.21.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.22
blk.22.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.22.attn_k_norm.weight
F32
F32
[128]
blk.22.attn_norm.weight
F32
F32
[2048]
blk.22.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.22.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.22.attn_q_norm.weight
F32
F32
[128]
blk.22.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.22.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.22.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.22.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.22.ffn_norm.weight
F32
F32
[2048]
blk.22.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.23
blk.23.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.23.attn_k_norm.weight
F32
F32
[128]
blk.23.attn_norm.weight
F32
F32
[2048]
blk.23.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.23.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.23.attn_q_norm.weight
F32
F32
[128]
blk.23.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.23.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.23.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.23.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.23.ffn_norm.weight
F32
F32
[2048]
blk.23.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.24
blk.24.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.24.attn_k_norm.weight
F32
F32
[128]
blk.24.attn_norm.weight
F32
F32
[2048]
blk.24.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.24.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.24.attn_q_norm.weight
F32
F32
[128]
blk.24.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.24.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.24.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.24.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.24.ffn_norm.weight
F32
F32
[2048]
blk.24.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.25
blk.25.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.25.attn_k_norm.weight
F32
F32
[128]
blk.25.attn_norm.weight
F32
F32
[2048]
blk.25.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.25.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.25.attn_q_norm.weight
F32
F32
[128]
blk.25.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.25.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.25.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.25.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.25.ffn_norm.weight
F32
F32
[2048]
blk.25.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.26
blk.26.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.26.attn_k_norm.weight
F32
F32
[128]
blk.26.attn_norm.weight
F32
F32
[2048]
blk.26.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.26.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.26.attn_q_norm.weight
F32
F32
[128]
blk.26.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.26.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.26.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.26.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.26.ffn_norm.weight
F32
F32
[2048]
blk.26.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.27
blk.27.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.27.attn_k_norm.weight
F32
F32
[128]
blk.27.attn_norm.weight
F32
F32
[2048]
blk.27.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.27.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.27.attn_q_norm.weight
F32
F32
[128]
blk.27.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.27.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.27.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.27.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.27.ffn_norm.weight
F32
F32
[2048]
blk.27.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.28
blk.28.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.28.attn_k_norm.weight
F32
F32
[128]
blk.28.attn_norm.weight
F32
F32
[2048]
blk.28.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.28.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.28.attn_q_norm.weight
F32
F32
[128]
blk.28.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.28.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.28.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.28.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.28.ffn_norm.weight
F32
F32
[2048]
blk.28.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.29
blk.29.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.29.attn_k_norm.weight
F32
F32
[128]
blk.29.attn_norm.weight
F32
F32
[2048]
blk.29.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.29.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.29.attn_q_norm.weight
F32
F32
[128]
blk.29.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.29.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.29.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.29.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.29.ffn_norm.weight
F32
F32
[2048]
blk.29.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.30
blk.30.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.30.attn_k_norm.weight
F32
F32
[128]
blk.30.attn_norm.weight
F32
F32
[2048]
blk.30.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.30.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.30.attn_q_norm.weight
F32
F32
[128]
blk.30.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.30.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.30.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.30.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.30.ffn_norm.weight
F32
F32
[2048]
blk.30.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.31
blk.31.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.31.attn_k_norm.weight
F32
F32
[128]
blk.31.attn_norm.weight
F32
F32
[2048]
blk.31.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.31.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.31.attn_q_norm.weight
F32
F32
[128]
blk.31.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.31.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.31.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.31.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.31.ffn_norm.weight
F32
F32
[2048]
blk.31.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.32
blk.32.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.32.attn_k_norm.weight
F32
F32
[128]
blk.32.attn_norm.weight
F32
F32
[2048]
blk.32.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.32.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.32.attn_q_norm.weight
F32
F32
[128]
blk.32.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.32.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.32.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.32.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.32.ffn_norm.weight
F32
F32
[2048]
blk.32.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.33
blk.33.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.33.attn_k_norm.weight
F32
F32
[128]
blk.33.attn_norm.weight
F32
F32
[2048]
blk.33.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.33.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.33.attn_q_norm.weight
F32
F32
[128]
blk.33.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.33.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.33.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.33.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.33.ffn_norm.weight
F32
F32
[2048]
blk.33.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.34
blk.34.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.34.attn_k_norm.weight
F32
F32
[128]
blk.34.attn_norm.weight
F32
F32
[2048]
blk.34.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.34.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.34.attn_q_norm.weight
F32
F32
[128]
blk.34.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.34.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.34.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.34.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.34.ffn_norm.weight
F32
F32
[2048]
blk.34.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.35
blk.35.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.35.attn_k_norm.weight
F32
F32
[128]
blk.35.attn_norm.weight
F32
F32
[2048]
blk.35.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.35.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.35.attn_q_norm.weight
F32
F32
[128]
blk.35.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.35.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.35.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.35.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.35.ffn_norm.weight
F32
F32
[2048]
blk.35.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.36
blk.36.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.36.attn_k_norm.weight
F32
F32
[128]
blk.36.attn_norm.weight
F32
F32
[2048]
blk.36.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.36.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.36.attn_q_norm.weight
F32
F32
[128]
blk.36.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.36.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.36.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.36.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.36.ffn_norm.weight
F32
F32
[2048]
blk.36.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.37
blk.37.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.37.attn_k_norm.weight
F32
F32
[128]
blk.37.attn_norm.weight
F32
F32
[2048]
blk.37.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.37.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.37.attn_q_norm.weight
F32
F32
[128]
blk.37.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.37.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.37.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.37.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.37.ffn_norm.weight
F32
F32
[2048]
blk.37.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.38
blk.38.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.38.attn_k_norm.weight
F32
F32
[128]
blk.38.attn_norm.weight
F32
F32
[2048]
blk.38.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.38.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.38.attn_q_norm.weight
F32
F32
[128]
blk.38.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.38.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.38.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.38.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.38.ffn_norm.weight
F32
F32
[2048]
blk.38.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.39
blk.39.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.39.attn_k_norm.weight
F32
F32
[128]
blk.39.attn_norm.weight
F32
F32
[2048]
blk.39.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.39.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.39.attn_q_norm.weight
F32
F32
[128]
blk.39.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.39.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.39.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.39.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.39.ffn_norm.weight
F32
F32
[2048]
blk.39.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.40
blk.40.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.40.attn_k_norm.weight
F32
F32
[128]
blk.40.attn_norm.weight
F32
F32
[2048]
blk.40.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.40.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.40.attn_q_norm.weight
F32
F32
[128]
blk.40.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.40.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.40.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.40.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.40.ffn_norm.weight
F32
F32
[2048]
blk.40.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.41
blk.41.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.41.attn_k_norm.weight
F32
F32
[128]
blk.41.attn_norm.weight
F32
F32
[2048]
blk.41.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.41.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.41.attn_q_norm.weight
F32
F32
[128]
blk.41.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.41.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.41.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.41.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.41.ffn_norm.weight
F32
F32
[2048]
blk.41.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.42
blk.42.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.42.attn_k_norm.weight
F32
F32
[128]
blk.42.attn_norm.weight
F32
F32
[2048]
blk.42.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.42.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.42.attn_q_norm.weight
F32
F32
[128]
blk.42.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.42.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.42.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.42.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.42.ffn_norm.weight
F32
F32
[2048]
blk.42.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.43
blk.43.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.43.attn_k_norm.weight
F32
F32
[128]
blk.43.attn_norm.weight
F32
F32
[2048]
blk.43.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.43.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.43.attn_q_norm.weight
F32
F32
[128]
blk.43.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.43.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.43.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.43.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.43.ffn_norm.weight
F32
F32
[2048]
blk.43.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.44
blk.44.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.44.attn_k_norm.weight
F32
F32
[128]
blk.44.attn_norm.weight
F32
F32
[2048]
blk.44.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.44.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.44.attn_q_norm.weight
F32
F32
[128]
blk.44.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.44.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.44.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.44.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.44.ffn_norm.weight
F32
F32
[2048]
blk.44.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.45
blk.45.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.45.attn_k_norm.weight
F32
F32
[128]
blk.45.attn_norm.weight
F32
F32
[2048]
blk.45.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.45.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.45.attn_q_norm.weight
F32
F32
[128]
blk.45.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.45.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.45.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.45.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.45.ffn_norm.weight
F32
F32
[2048]
blk.45.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.46
blk.46.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.46.attn_k_norm.weight
F32
F32
[128]
blk.46.attn_norm.weight
F32
F32
[2048]
blk.46.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.46.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.46.attn_q_norm.weight
F32
F32
[128]
blk.46.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.46.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.46.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.46.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.46.ffn_norm.weight
F32
F32
[2048]
blk.46.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.47
blk.47.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.47.attn_k_norm.weight
F32
F32
[128]
blk.47.attn_norm.weight
F32
F32
[2048]
blk.47.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.47.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.47.attn_q_norm.weight
F32
F32
[128]
blk.47.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.47.ffn_down_exps.weight
Q8_0
Q8_0
[768, 2048, 128]
blk.47.ffn_gate_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
blk.47.ffn_gate_inp.weight
F32
F32
[2048, 128]
blk.47.ffn_norm.weight
F32
F32
[2048]
blk.47.ffn_up_exps.weight
Q8_0
Q8_0
[2048, 768, 128]
output.weight
Q8_0
Q8_0
[2048, 151936]
output_norm.weight
F32
F32
[2048]