Models
GitHub
Discord
Turbo
Sign in
Download
Models
Download
GitHub
Discord
Sign in
alibayram
/
doktor-gemma3-27b
:latest
2
Downloads
Updated
2 months ago
Sağlık alanında uzmanlaşmış büyük bir dil modelidir. Bu model, hasta sorularına daha doğru ve etkili yanıtlar vermek üzere tasarlanmış ve eğitilmiştir.
Sağlık alanında uzmanlaşmış büyük bir dil modelidir. Bu model, hasta sorularına daha doğru ve etkili yanıtlar vermek üzere tasarlanmış ve eğitilmiştir.
Cancel
doktor-gemma3-27b:latest
...
/
model
4f8a2030a5b0 · 29GB
Metadata
general.architecture
gemma3
gemma3
general.file_type
Q8_0
Q8_0
gemma3.attention.head_count
32
32
gemma3.attention.head_count_kv
16
16
gemma3.attention.key_length
128
128
gemma3.attention.layer_norm_rms_epsilon
1e-06
1e-06
gemma3.attention.sliding_window
1024
1024
gemma3.attention.value_length
128
128
gemma3.block_count
62
62
gemma3.context_length
131072
131072
gemma3.embedding_length
5376
5376
gemma3.feed_forward_length
21504
21504
gemma3.rope.freq_base
1e+06
1e+06
gemma3.rope.scaling.factor
8
8
gemma3.rope.scaling.type
linear
linear
tokenizer.ggml.add_bos_token
true
true
tokenizer.ggml.add_eos_token
false
false
tokenizer.ggml.add_sep_token
false
false
tokenizer.ggml.add_space_prefix
false
false
tokenizer.ggml.bos_token_id
2
2
tokenizer.ggml.eos_token_id
106
106
tokenizer.ggml.model
llama
llama
tokenizer.ggml.padding_token_id
0
0
tokenizer.ggml.pre
default
default
tokenizer.ggml.scores
[-1000, -1000, -1000, -1000, -1000, ...]
[-1000, -1000, -1000, -1000, -1000, ...]
tokenizer.ggml.token_type
[3, 3, 3, 3, 3, ...]
[3, 3, 3, 3, 3, ...]
tokenizer.ggml.tokens
[<pad>, <eos>, <bos>, <unk>, <mask>, ...]
[<pad>, <eos>, <bos>, <unk>, <mask>, ...]
tokenizer.ggml.unknown_token_id
3
3
Tensor
Name
Type
Shape
token_embd.weight
Q8_0
Q8_0
[5376, 262208]
blk.0
blk.0.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.0.attn_k_norm.weight
F32
F32
[128]
blk.0.attn_norm.weight
F32
F32
[5376]
blk.0.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.0.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.0.attn_q_norm.weight
F32
F32
[128]
blk.0.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.0.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.0.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.0.ffn_norm.weight
F32
F32
[5376]
blk.0.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.0.post_attention_norm.weight
F32
F32
[5376]
blk.0.post_ffw_norm.weight
F32
F32
[5376]
blk.1
blk.1.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.1.attn_k_norm.weight
F32
F32
[128]
blk.1.attn_norm.weight
F32
F32
[5376]
blk.1.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.1.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.1.attn_q_norm.weight
F32
F32
[128]
blk.1.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.1.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.1.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.1.ffn_norm.weight
F32
F32
[5376]
blk.1.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.1.post_attention_norm.weight
F32
F32
[5376]
blk.1.post_ffw_norm.weight
F32
F32
[5376]
blk.2
blk.2.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.2.attn_k_norm.weight
F32
F32
[128]
blk.2.attn_norm.weight
F32
F32
[5376]
blk.2.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.2.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.2.attn_q_norm.weight
F32
F32
[128]
blk.2.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.2.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.2.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.2.ffn_norm.weight
F32
F32
[5376]
blk.2.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.2.post_attention_norm.weight
F32
F32
[5376]
blk.2.post_ffw_norm.weight
F32
F32
[5376]
blk.3
blk.3.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.3.attn_k_norm.weight
F32
F32
[128]
blk.3.attn_norm.weight
F32
F32
[5376]
blk.3.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.3.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.3.attn_q_norm.weight
F32
F32
[128]
blk.3.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.3.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.3.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.3.ffn_norm.weight
F32
F32
[5376]
blk.3.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.3.post_attention_norm.weight
F32
F32
[5376]
blk.3.post_ffw_norm.weight
F32
F32
[5376]
blk.4
blk.4.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.4.attn_k_norm.weight
F32
F32
[128]
blk.4.attn_norm.weight
F32
F32
[5376]
blk.4.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.4.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.4.attn_q_norm.weight
F32
F32
[128]
blk.4.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.4.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.4.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.4.ffn_norm.weight
F32
F32
[5376]
blk.4.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.4.post_attention_norm.weight
F32
F32
[5376]
blk.4.post_ffw_norm.weight
F32
F32
[5376]
blk.5
blk.5.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.5.attn_k_norm.weight
F32
F32
[128]
blk.5.attn_norm.weight
F32
F32
[5376]
blk.5.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.5.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.5.attn_q_norm.weight
F32
F32
[128]
blk.5.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.5.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.5.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.5.ffn_norm.weight
F32
F32
[5376]
blk.5.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.5.post_attention_norm.weight
F32
F32
[5376]
blk.5.post_ffw_norm.weight
F32
F32
[5376]
blk.6
blk.6.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.6.attn_k_norm.weight
F32
F32
[128]
blk.6.attn_norm.weight
F32
F32
[5376]
blk.6.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.6.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.6.attn_q_norm.weight
F32
F32
[128]
blk.6.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.6.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.6.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.6.ffn_norm.weight
F32
F32
[5376]
blk.6.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.6.post_attention_norm.weight
F32
F32
[5376]
blk.6.post_ffw_norm.weight
F32
F32
[5376]
blk.7
blk.7.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.7.attn_k_norm.weight
F32
F32
[128]
blk.7.attn_norm.weight
F32
F32
[5376]
blk.7.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.7.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.7.attn_q_norm.weight
F32
F32
[128]
blk.7.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.7.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.7.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.7.ffn_norm.weight
F32
F32
[5376]
blk.7.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.7.post_attention_norm.weight
F32
F32
[5376]
blk.7.post_ffw_norm.weight
F32
F32
[5376]
blk.8
blk.8.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.8.attn_k_norm.weight
F32
F32
[128]
blk.8.attn_norm.weight
F32
F32
[5376]
blk.8.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.8.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.8.attn_q_norm.weight
F32
F32
[128]
blk.8.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.8.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.8.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.8.ffn_norm.weight
F32
F32
[5376]
blk.8.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.8.post_attention_norm.weight
F32
F32
[5376]
blk.8.post_ffw_norm.weight
F32
F32
[5376]
blk.9
blk.9.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.9.attn_k_norm.weight
F32
F32
[128]
blk.9.attn_norm.weight
F32
F32
[5376]
blk.9.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.9.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.9.attn_q_norm.weight
F32
F32
[128]
blk.9.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.9.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.9.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.9.ffn_norm.weight
F32
F32
[5376]
blk.9.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.9.post_attention_norm.weight
F32
F32
[5376]
blk.9.post_ffw_norm.weight
F32
F32
[5376]
blk.10
blk.10.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.10.attn_k_norm.weight
F32
F32
[128]
blk.10.attn_norm.weight
F32
F32
[5376]
blk.10.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.10.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.10.attn_q_norm.weight
F32
F32
[128]
blk.10.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.10.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.10.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.10.ffn_norm.weight
F32
F32
[5376]
blk.10.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.10.post_attention_norm.weight
F32
F32
[5376]
blk.10.post_ffw_norm.weight
F32
F32
[5376]
blk.11
blk.11.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.11.attn_k_norm.weight
F32
F32
[128]
blk.11.attn_norm.weight
F32
F32
[5376]
blk.11.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.11.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.11.attn_q_norm.weight
F32
F32
[128]
blk.11.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.11.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.11.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.11.ffn_norm.weight
F32
F32
[5376]
blk.11.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.11.post_attention_norm.weight
F32
F32
[5376]
blk.11.post_ffw_norm.weight
F32
F32
[5376]
blk.12
blk.12.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.12.attn_k_norm.weight
F32
F32
[128]
blk.12.attn_norm.weight
F32
F32
[5376]
blk.12.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.12.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.12.attn_q_norm.weight
F32
F32
[128]
blk.12.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.12.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.12.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.12.ffn_norm.weight
F32
F32
[5376]
blk.12.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.12.post_attention_norm.weight
F32
F32
[5376]
blk.12.post_ffw_norm.weight
F32
F32
[5376]
blk.13
blk.13.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.13.attn_k_norm.weight
F32
F32
[128]
blk.13.attn_norm.weight
F32
F32
[5376]
blk.13.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.13.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.13.attn_q_norm.weight
F32
F32
[128]
blk.13.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.13.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.13.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.13.ffn_norm.weight
F32
F32
[5376]
blk.13.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.13.post_attention_norm.weight
F32
F32
[5376]
blk.13.post_ffw_norm.weight
F32
F32
[5376]
blk.14
blk.14.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.14.attn_k_norm.weight
F32
F32
[128]
blk.14.attn_norm.weight
F32
F32
[5376]
blk.14.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.14.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.14.attn_q_norm.weight
F32
F32
[128]
blk.14.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.14.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.14.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.14.ffn_norm.weight
F32
F32
[5376]
blk.14.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.14.post_attention_norm.weight
F32
F32
[5376]
blk.14.post_ffw_norm.weight
F32
F32
[5376]
blk.15
blk.15.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.15.attn_k_norm.weight
F32
F32
[128]
blk.15.attn_norm.weight
F32
F32
[5376]
blk.15.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.15.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.15.attn_q_norm.weight
F32
F32
[128]
blk.15.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.15.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.15.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.15.ffn_norm.weight
F32
F32
[5376]
blk.15.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.15.post_attention_norm.weight
F32
F32
[5376]
blk.15.post_ffw_norm.weight
F32
F32
[5376]
blk.16
blk.16.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.16.attn_k_norm.weight
F32
F32
[128]
blk.16.attn_norm.weight
F32
F32
[5376]
blk.16.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.16.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.16.attn_q_norm.weight
F32
F32
[128]
blk.16.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.16.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.16.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.16.ffn_norm.weight
F32
F32
[5376]
blk.16.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.16.post_attention_norm.weight
F32
F32
[5376]
blk.16.post_ffw_norm.weight
F32
F32
[5376]
blk.17
blk.17.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.17.attn_k_norm.weight
F32
F32
[128]
blk.17.attn_norm.weight
F32
F32
[5376]
blk.17.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.17.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.17.attn_q_norm.weight
F32
F32
[128]
blk.17.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.17.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.17.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.17.ffn_norm.weight
F32
F32
[5376]
blk.17.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.17.post_attention_norm.weight
F32
F32
[5376]
blk.17.post_ffw_norm.weight
F32
F32
[5376]
blk.18
blk.18.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.18.attn_k_norm.weight
F32
F32
[128]
blk.18.attn_norm.weight
F32
F32
[5376]
blk.18.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.18.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.18.attn_q_norm.weight
F32
F32
[128]
blk.18.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.18.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.18.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.18.ffn_norm.weight
F32
F32
[5376]
blk.18.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.18.post_attention_norm.weight
F32
F32
[5376]
blk.18.post_ffw_norm.weight
F32
F32
[5376]
blk.19
blk.19.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.19.attn_k_norm.weight
F32
F32
[128]
blk.19.attn_norm.weight
F32
F32
[5376]
blk.19.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.19.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.19.attn_q_norm.weight
F32
F32
[128]
blk.19.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.19.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.19.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.19.ffn_norm.weight
F32
F32
[5376]
blk.19.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.19.post_attention_norm.weight
F32
F32
[5376]
blk.19.post_ffw_norm.weight
F32
F32
[5376]
blk.20
blk.20.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.20.attn_k_norm.weight
F32
F32
[128]
blk.20.attn_norm.weight
F32
F32
[5376]
blk.20.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.20.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.20.attn_q_norm.weight
F32
F32
[128]
blk.20.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.20.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.20.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.20.ffn_norm.weight
F32
F32
[5376]
blk.20.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.20.post_attention_norm.weight
F32
F32
[5376]
blk.20.post_ffw_norm.weight
F32
F32
[5376]
blk.21
blk.21.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.21.attn_k_norm.weight
F32
F32
[128]
blk.21.attn_norm.weight
F32
F32
[5376]
blk.21.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.21.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.21.attn_q_norm.weight
F32
F32
[128]
blk.21.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.21.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.21.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.21.ffn_norm.weight
F32
F32
[5376]
blk.21.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.21.post_attention_norm.weight
F32
F32
[5376]
blk.21.post_ffw_norm.weight
F32
F32
[5376]
blk.22
blk.22.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.22.attn_k_norm.weight
F32
F32
[128]
blk.22.attn_norm.weight
F32
F32
[5376]
blk.22.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.22.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.22.attn_q_norm.weight
F32
F32
[128]
blk.22.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.22.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.22.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.22.ffn_norm.weight
F32
F32
[5376]
blk.22.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.22.post_attention_norm.weight
F32
F32
[5376]
blk.22.post_ffw_norm.weight
F32
F32
[5376]
blk.23
blk.23.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.23.attn_k_norm.weight
F32
F32
[128]
blk.23.attn_norm.weight
F32
F32
[5376]
blk.23.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.23.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.23.attn_q_norm.weight
F32
F32
[128]
blk.23.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.23.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.23.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.23.ffn_norm.weight
F32
F32
[5376]
blk.23.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.23.post_attention_norm.weight
F32
F32
[5376]
blk.23.post_ffw_norm.weight
F32
F32
[5376]
blk.24
blk.24.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.24.attn_k_norm.weight
F32
F32
[128]
blk.24.attn_norm.weight
F32
F32
[5376]
blk.24.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.24.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.24.attn_q_norm.weight
F32
F32
[128]
blk.24.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.24.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.24.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.24.ffn_norm.weight
F32
F32
[5376]
blk.24.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.24.post_attention_norm.weight
F32
F32
[5376]
blk.24.post_ffw_norm.weight
F32
F32
[5376]
blk.25
blk.25.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.25.attn_k_norm.weight
F32
F32
[128]
blk.25.attn_norm.weight
F32
F32
[5376]
blk.25.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.25.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.25.attn_q_norm.weight
F32
F32
[128]
blk.25.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.25.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.25.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.25.ffn_norm.weight
F32
F32
[5376]
blk.25.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.25.post_attention_norm.weight
F32
F32
[5376]
blk.25.post_ffw_norm.weight
F32
F32
[5376]
blk.26
blk.26.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.26.attn_k_norm.weight
F32
F32
[128]
blk.26.attn_norm.weight
F32
F32
[5376]
blk.26.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.26.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.26.attn_q_norm.weight
F32
F32
[128]
blk.26.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.26.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.26.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.26.ffn_norm.weight
F32
F32
[5376]
blk.26.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.26.post_attention_norm.weight
F32
F32
[5376]
blk.26.post_ffw_norm.weight
F32
F32
[5376]
blk.27
blk.27.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.27.attn_k_norm.weight
F32
F32
[128]
blk.27.attn_norm.weight
F32
F32
[5376]
blk.27.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.27.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.27.attn_q_norm.weight
F32
F32
[128]
blk.27.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.27.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.27.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.27.ffn_norm.weight
F32
F32
[5376]
blk.27.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.27.post_attention_norm.weight
F32
F32
[5376]
blk.27.post_ffw_norm.weight
F32
F32
[5376]
blk.28
blk.28.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.28.attn_k_norm.weight
F32
F32
[128]
blk.28.attn_norm.weight
F32
F32
[5376]
blk.28.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.28.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.28.attn_q_norm.weight
F32
F32
[128]
blk.28.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.28.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.28.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.28.ffn_norm.weight
F32
F32
[5376]
blk.28.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.28.post_attention_norm.weight
F32
F32
[5376]
blk.28.post_ffw_norm.weight
F32
F32
[5376]
blk.29
blk.29.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.29.attn_k_norm.weight
F32
F32
[128]
blk.29.attn_norm.weight
F32
F32
[5376]
blk.29.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.29.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.29.attn_q_norm.weight
F32
F32
[128]
blk.29.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.29.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.29.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.29.ffn_norm.weight
F32
F32
[5376]
blk.29.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.29.post_attention_norm.weight
F32
F32
[5376]
blk.29.post_ffw_norm.weight
F32
F32
[5376]
blk.30
blk.30.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.30.attn_k_norm.weight
F32
F32
[128]
blk.30.attn_norm.weight
F32
F32
[5376]
blk.30.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.30.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.30.attn_q_norm.weight
F32
F32
[128]
blk.30.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.30.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.30.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.30.ffn_norm.weight
F32
F32
[5376]
blk.30.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.30.post_attention_norm.weight
F32
F32
[5376]
blk.30.post_ffw_norm.weight
F32
F32
[5376]
blk.31
blk.31.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.31.attn_k_norm.weight
F32
F32
[128]
blk.31.attn_norm.weight
F32
F32
[5376]
blk.31.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.31.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.31.attn_q_norm.weight
F32
F32
[128]
blk.31.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.31.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.31.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.31.ffn_norm.weight
F32
F32
[5376]
blk.31.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.31.post_attention_norm.weight
F32
F32
[5376]
blk.31.post_ffw_norm.weight
F32
F32
[5376]
blk.32
blk.32.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.32.attn_k_norm.weight
F32
F32
[128]
blk.32.attn_norm.weight
F32
F32
[5376]
blk.32.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.32.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.32.attn_q_norm.weight
F32
F32
[128]
blk.32.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.32.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.32.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.32.ffn_norm.weight
F32
F32
[5376]
blk.32.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.32.post_attention_norm.weight
F32
F32
[5376]
blk.32.post_ffw_norm.weight
F32
F32
[5376]
blk.33
blk.33.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.33.attn_k_norm.weight
F32
F32
[128]
blk.33.attn_norm.weight
F32
F32
[5376]
blk.33.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.33.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.33.attn_q_norm.weight
F32
F32
[128]
blk.33.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.33.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.33.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.33.ffn_norm.weight
F32
F32
[5376]
blk.33.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.33.post_attention_norm.weight
F32
F32
[5376]
blk.33.post_ffw_norm.weight
F32
F32
[5376]
blk.34
blk.34.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.34.attn_k_norm.weight
F32
F32
[128]
blk.34.attn_norm.weight
F32
F32
[5376]
blk.34.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.34.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.34.attn_q_norm.weight
F32
F32
[128]
blk.34.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.34.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.34.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.34.ffn_norm.weight
F32
F32
[5376]
blk.34.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.34.post_attention_norm.weight
F32
F32
[5376]
blk.34.post_ffw_norm.weight
F32
F32
[5376]
blk.35
blk.35.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.35.attn_k_norm.weight
F32
F32
[128]
blk.35.attn_norm.weight
F32
F32
[5376]
blk.35.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.35.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.35.attn_q_norm.weight
F32
F32
[128]
blk.35.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.35.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.35.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.35.ffn_norm.weight
F32
F32
[5376]
blk.35.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.35.post_attention_norm.weight
F32
F32
[5376]
blk.35.post_ffw_norm.weight
F32
F32
[5376]
blk.36
blk.36.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.36.attn_k_norm.weight
F32
F32
[128]
blk.36.attn_norm.weight
F32
F32
[5376]
blk.36.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.36.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.36.attn_q_norm.weight
F32
F32
[128]
blk.36.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.36.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.36.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.36.ffn_norm.weight
F32
F32
[5376]
blk.36.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.36.post_attention_norm.weight
F32
F32
[5376]
blk.36.post_ffw_norm.weight
F32
F32
[5376]
blk.37
blk.37.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.37.attn_k_norm.weight
F32
F32
[128]
blk.37.attn_norm.weight
F32
F32
[5376]
blk.37.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.37.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.37.attn_q_norm.weight
F32
F32
[128]
blk.37.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.37.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.37.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.37.ffn_norm.weight
F32
F32
[5376]
blk.37.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.37.post_attention_norm.weight
F32
F32
[5376]
blk.37.post_ffw_norm.weight
F32
F32
[5376]
blk.38
blk.38.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.38.attn_k_norm.weight
F32
F32
[128]
blk.38.attn_norm.weight
F32
F32
[5376]
blk.38.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.38.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.38.attn_q_norm.weight
F32
F32
[128]
blk.38.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.38.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.38.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.38.ffn_norm.weight
F32
F32
[5376]
blk.38.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.38.post_attention_norm.weight
F32
F32
[5376]
blk.38.post_ffw_norm.weight
F32
F32
[5376]
blk.39
blk.39.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.39.attn_k_norm.weight
F32
F32
[128]
blk.39.attn_norm.weight
F32
F32
[5376]
blk.39.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.39.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.39.attn_q_norm.weight
F32
F32
[128]
blk.39.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.39.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.39.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.39.ffn_norm.weight
F32
F32
[5376]
blk.39.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.39.post_attention_norm.weight
F32
F32
[5376]
blk.39.post_ffw_norm.weight
F32
F32
[5376]
blk.40
blk.40.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.40.attn_k_norm.weight
F32
F32
[128]
blk.40.attn_norm.weight
F32
F32
[5376]
blk.40.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.40.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.40.attn_q_norm.weight
F32
F32
[128]
blk.40.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.40.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.40.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.40.ffn_norm.weight
F32
F32
[5376]
blk.40.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.40.post_attention_norm.weight
F32
F32
[5376]
blk.40.post_ffw_norm.weight
F32
F32
[5376]
blk.41
blk.41.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.41.attn_k_norm.weight
F32
F32
[128]
blk.41.attn_norm.weight
F32
F32
[5376]
blk.41.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.41.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.41.attn_q_norm.weight
F32
F32
[128]
blk.41.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.41.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.41.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.41.ffn_norm.weight
F32
F32
[5376]
blk.41.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.41.post_attention_norm.weight
F32
F32
[5376]
blk.41.post_ffw_norm.weight
F32
F32
[5376]
blk.42
blk.42.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.42.attn_k_norm.weight
F32
F32
[128]
blk.42.attn_norm.weight
F32
F32
[5376]
blk.42.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.42.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.42.attn_q_norm.weight
F32
F32
[128]
blk.42.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.42.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.42.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.42.ffn_norm.weight
F32
F32
[5376]
blk.42.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.42.post_attention_norm.weight
F32
F32
[5376]
blk.42.post_ffw_norm.weight
F32
F32
[5376]
blk.43
blk.43.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.43.attn_k_norm.weight
F32
F32
[128]
blk.43.attn_norm.weight
F32
F32
[5376]
blk.43.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.43.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.43.attn_q_norm.weight
F32
F32
[128]
blk.43.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.43.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.43.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.43.ffn_norm.weight
F32
F32
[5376]
blk.43.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.43.post_attention_norm.weight
F32
F32
[5376]
blk.43.post_ffw_norm.weight
F32
F32
[5376]
blk.44
blk.44.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.44.attn_k_norm.weight
F32
F32
[128]
blk.44.attn_norm.weight
F32
F32
[5376]
blk.44.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.44.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.44.attn_q_norm.weight
F32
F32
[128]
blk.44.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.44.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.44.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.44.ffn_norm.weight
F32
F32
[5376]
blk.44.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.44.post_attention_norm.weight
F32
F32
[5376]
blk.44.post_ffw_norm.weight
F32
F32
[5376]
blk.45
blk.45.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.45.attn_k_norm.weight
F32
F32
[128]
blk.45.attn_norm.weight
F32
F32
[5376]
blk.45.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.45.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.45.attn_q_norm.weight
F32
F32
[128]
blk.45.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.45.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.45.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.45.ffn_norm.weight
F32
F32
[5376]
blk.45.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.45.post_attention_norm.weight
F32
F32
[5376]
blk.45.post_ffw_norm.weight
F32
F32
[5376]
blk.46
blk.46.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.46.attn_k_norm.weight
F32
F32
[128]
blk.46.attn_norm.weight
F32
F32
[5376]
blk.46.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.46.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.46.attn_q_norm.weight
F32
F32
[128]
blk.46.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.46.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.46.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.46.ffn_norm.weight
F32
F32
[5376]
blk.46.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.46.post_attention_norm.weight
F32
F32
[5376]
blk.46.post_ffw_norm.weight
F32
F32
[5376]
blk.47
blk.47.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.47.attn_k_norm.weight
F32
F32
[128]
blk.47.attn_norm.weight
F32
F32
[5376]
blk.47.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.47.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.47.attn_q_norm.weight
F32
F32
[128]
blk.47.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.47.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.47.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.47.ffn_norm.weight
F32
F32
[5376]
blk.47.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.47.post_attention_norm.weight
F32
F32
[5376]
blk.47.post_ffw_norm.weight
F32
F32
[5376]
blk.48
blk.48.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.48.attn_k_norm.weight
F32
F32
[128]
blk.48.attn_norm.weight
F32
F32
[5376]
blk.48.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.48.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.48.attn_q_norm.weight
F32
F32
[128]
blk.48.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.48.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.48.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.48.ffn_norm.weight
F32
F32
[5376]
blk.48.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.48.post_attention_norm.weight
F32
F32
[5376]
blk.48.post_ffw_norm.weight
F32
F32
[5376]
blk.49
blk.49.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.49.attn_k_norm.weight
F32
F32
[128]
blk.49.attn_norm.weight
F32
F32
[5376]
blk.49.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.49.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.49.attn_q_norm.weight
F32
F32
[128]
blk.49.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.49.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.49.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.49.ffn_norm.weight
F32
F32
[5376]
blk.49.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.49.post_attention_norm.weight
F32
F32
[5376]
blk.49.post_ffw_norm.weight
F32
F32
[5376]
blk.50
blk.50.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.50.attn_k_norm.weight
F32
F32
[128]
blk.50.attn_norm.weight
F32
F32
[5376]
blk.50.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.50.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.50.attn_q_norm.weight
F32
F32
[128]
blk.50.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.50.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.50.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.50.ffn_norm.weight
F32
F32
[5376]
blk.50.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.50.post_attention_norm.weight
F32
F32
[5376]
blk.50.post_ffw_norm.weight
F32
F32
[5376]
blk.51
blk.51.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.51.attn_k_norm.weight
F32
F32
[128]
blk.51.attn_norm.weight
F32
F32
[5376]
blk.51.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.51.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.51.attn_q_norm.weight
F32
F32
[128]
blk.51.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.51.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.51.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.51.ffn_norm.weight
F32
F32
[5376]
blk.51.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.51.post_attention_norm.weight
F32
F32
[5376]
blk.51.post_ffw_norm.weight
F32
F32
[5376]
blk.52
blk.52.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.52.attn_k_norm.weight
F32
F32
[128]
blk.52.attn_norm.weight
F32
F32
[5376]
blk.52.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.52.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.52.attn_q_norm.weight
F32
F32
[128]
blk.52.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.52.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.52.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.52.ffn_norm.weight
F32
F32
[5376]
blk.52.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.52.post_attention_norm.weight
F32
F32
[5376]
blk.52.post_ffw_norm.weight
F32
F32
[5376]
blk.53
blk.53.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.53.attn_k_norm.weight
F32
F32
[128]
blk.53.attn_norm.weight
F32
F32
[5376]
blk.53.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.53.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.53.attn_q_norm.weight
F32
F32
[128]
blk.53.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.53.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.53.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.53.ffn_norm.weight
F32
F32
[5376]
blk.53.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.53.post_attention_norm.weight
F32
F32
[5376]
blk.53.post_ffw_norm.weight
F32
F32
[5376]
blk.54
blk.54.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.54.attn_k_norm.weight
F32
F32
[128]
blk.54.attn_norm.weight
F32
F32
[5376]
blk.54.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.54.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.54.attn_q_norm.weight
F32
F32
[128]
blk.54.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.54.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.54.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.54.ffn_norm.weight
F32
F32
[5376]
blk.54.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.54.post_attention_norm.weight
F32
F32
[5376]
blk.54.post_ffw_norm.weight
F32
F32
[5376]
blk.55
blk.55.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.55.attn_k_norm.weight
F32
F32
[128]
blk.55.attn_norm.weight
F32
F32
[5376]
blk.55.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.55.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.55.attn_q_norm.weight
F32
F32
[128]
blk.55.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.55.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.55.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.55.ffn_norm.weight
F32
F32
[5376]
blk.55.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.55.post_attention_norm.weight
F32
F32
[5376]
blk.55.post_ffw_norm.weight
F32
F32
[5376]
blk.56
blk.56.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.56.attn_k_norm.weight
F32
F32
[128]
blk.56.attn_norm.weight
F32
F32
[5376]
blk.56.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.56.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.56.attn_q_norm.weight
F32
F32
[128]
blk.56.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.56.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.56.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.56.ffn_norm.weight
F32
F32
[5376]
blk.56.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.56.post_attention_norm.weight
F32
F32
[5376]
blk.56.post_ffw_norm.weight
F32
F32
[5376]
blk.57
blk.57.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.57.attn_k_norm.weight
F32
F32
[128]
blk.57.attn_norm.weight
F32
F32
[5376]
blk.57.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.57.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.57.attn_q_norm.weight
F32
F32
[128]
blk.57.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.57.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.57.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.57.ffn_norm.weight
F32
F32
[5376]
blk.57.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.57.post_attention_norm.weight
F32
F32
[5376]
blk.57.post_ffw_norm.weight
F32
F32
[5376]
blk.58
blk.58.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.58.attn_k_norm.weight
F32
F32
[128]
blk.58.attn_norm.weight
F32
F32
[5376]
blk.58.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.58.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.58.attn_q_norm.weight
F32
F32
[128]
blk.58.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.58.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.58.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.58.ffn_norm.weight
F32
F32
[5376]
blk.58.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.58.post_attention_norm.weight
F32
F32
[5376]
blk.58.post_ffw_norm.weight
F32
F32
[5376]
blk.59
blk.59.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.59.attn_k_norm.weight
F32
F32
[128]
blk.59.attn_norm.weight
F32
F32
[5376]
blk.59.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.59.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.59.attn_q_norm.weight
F32
F32
[128]
blk.59.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.59.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.59.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.59.ffn_norm.weight
F32
F32
[5376]
blk.59.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.59.post_attention_norm.weight
F32
F32
[5376]
blk.59.post_ffw_norm.weight
F32
F32
[5376]
blk.60
blk.60.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.60.attn_k_norm.weight
F32
F32
[128]
blk.60.attn_norm.weight
F32
F32
[5376]
blk.60.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.60.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.60.attn_q_norm.weight
F32
F32
[128]
blk.60.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.60.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.60.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.60.ffn_norm.weight
F32
F32
[5376]
blk.60.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.60.post_attention_norm.weight
F32
F32
[5376]
blk.60.post_ffw_norm.weight
F32
F32
[5376]
blk.61
blk.61.attn_k.weight
Q8_0
Q8_0
[5376, 2048]
blk.61.attn_k_norm.weight
F32
F32
[128]
blk.61.attn_norm.weight
F32
F32
[5376]
blk.61.attn_output.weight
Q8_0
Q8_0
[4096, 5376]
blk.61.attn_q.weight
Q8_0
Q8_0
[5376, 4096]
blk.61.attn_q_norm.weight
F32
F32
[128]
blk.61.attn_v.weight
Q8_0
Q8_0
[5376, 2048]
blk.61.ffn_down.weight
Q8_0
Q8_0
[21504, 5376]
blk.61.ffn_gate.weight
Q8_0
Q8_0
[5376, 21504]
blk.61.ffn_norm.weight
F32
F32
[5376]
blk.61.ffn_up.weight
Q8_0
Q8_0
[5376, 21504]
blk.61.post_attention_norm.weight
F32
F32
[5376]
blk.61.post_ffw_norm.weight
F32
F32
[5376]
output_norm.weight
F32
F32
[5376]