Models
GitHub
Discord
Docs
Cloud
Sign in
Download
Models
Download
GitHub
Discord
Docs
Cloud
Sign in
MichelRosselli
/
minimax-m2
:tq1_0
175
Downloads
Updated
1 week ago
Cancel
minimax-m2:tq1_0
...
/
model
aaf82562594a · 56GB
Metadata
general.architecture
minimax-m2
minimax-m2
general.file_type
IQ1_S
IQ1_S
minimax-m2.attention.head_count
48
48
minimax-m2.attention.head_count_kv
8
8
minimax-m2.attention.key_length
128
128
minimax-m2.attention.layer_norm_rms_epsilon
1e-06
1e-06
minimax-m2.attention.value_length
128
128
minimax-m2.block_count
62
62
minimax-m2.context_length
196608
196608
minimax-m2.embedding_length
3072
3072
minimax-m2.expert_count
256
256
minimax-m2.expert_feed_forward_length
1536
1536
minimax-m2.expert_gating_func
2
2
minimax-m2.expert_used_count
8
8
minimax-m2.feed_forward_length
1536
1536
minimax-m2.rope.dimension_count
64
64
minimax-m2.rope.freq_base
5e+06
5e+06
tokenizer.ggml.add_bos_token
true
true
tokenizer.ggml.bos_token_id
200034
200034
tokenizer.ggml.eos_token_id
200020
200020
tokenizer.ggml.merges
[Ġ Ġ, Ġ t, Ġ a, i n, e r, ...]
[Ġ Ġ, Ġ t, Ġ a, i n, e r, ...]
tokenizer.ggml.model
gpt2
gpt2
tokenizer.ggml.padding_token_id
200004
200004
tokenizer.ggml.pre
minimax-m2
minimax-m2
tokenizer.ggml.token_type
[1, 1, 1, 1, 1, ...]
[1, 1, 1, 1, 1, ...]
tokenizer.ggml.tokens
[Ā, ā, Ă, ă, Ą, ...]
[Ā, ā, Ă, ă, Ą, ...]
tokenizer.ggml.unknown_token_id
200021
200021
quantize.imatrix.chunks_count
697
697
quantize.imatrix.dataset
unsloth_calibration_MiniMax-M2.txt
unsloth_calibration_MiniMax-M2.txt
quantize.imatrix.entries_count
496
496
quantize.imatrix.file
MiniMax-M2-GGUF/imatrix_unsloth.gguf
MiniMax-M2-GGUF/imatrix_unsloth.gguf
Tensor
Name
Type
Shape
token_embd.weight
Q4_K
Q4_K
[3072, 200064]
blk.0
blk.0.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.0.attn_k_norm.weight
F32
F32
[1024]
blk.0.attn_norm.weight
F32
F32
[3072]
blk.0.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.0.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.0.attn_q_norm.weight
F32
F32
[6144]
blk.0.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.0.exp_probs_b.bias
F32
F32
[256]
blk.0.ffn_down_exps.weight
IQ3_S
IQ3_S
[1536, 3072, 256]
blk.0.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.0.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.0.ffn_norm.weight
F32
F32
[3072]
blk.0.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.1
blk.1.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.1.attn_k_norm.weight
F32
F32
[1024]
blk.1.attn_norm.weight
F32
F32
[3072]
blk.1.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.1.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.1.attn_q_norm.weight
F32
F32
[6144]
blk.1.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.1.exp_probs_b.bias
F32
F32
[256]
blk.1.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.1.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.1.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.1.ffn_norm.weight
F32
F32
[3072]
blk.1.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.2
blk.2.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.2.attn_k_norm.weight
F32
F32
[1024]
blk.2.attn_norm.weight
F32
F32
[3072]
blk.2.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.2.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.2.attn_q_norm.weight
F32
F32
[6144]
blk.2.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.2.exp_probs_b.bias
F32
F32
[256]
blk.2.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.2.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.2.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.2.ffn_norm.weight
F32
F32
[3072]
blk.2.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.3
blk.3.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.3.attn_k_norm.weight
F32
F32
[1024]
blk.3.attn_norm.weight
F32
F32
[3072]
blk.3.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.3.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.3.attn_q_norm.weight
F32
F32
[6144]
blk.3.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.3.exp_probs_b.bias
F32
F32
[256]
blk.3.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.3.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.3.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.3.ffn_norm.weight
F32
F32
[3072]
blk.3.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.4
blk.4.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.4.attn_k_norm.weight
F32
F32
[1024]
blk.4.attn_norm.weight
F32
F32
[3072]
blk.4.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.4.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.4.attn_q_norm.weight
F32
F32
[6144]
blk.4.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.4.exp_probs_b.bias
F32
F32
[256]
blk.4.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.4.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.4.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.4.ffn_norm.weight
F32
F32
[3072]
blk.4.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.5
blk.5.attn_k.weight
Q5_K
Q5_K
[3072, 1024]
blk.5.attn_k_norm.weight
F32
F32
[1024]
blk.5.attn_norm.weight
F32
F32
[3072]
blk.5.attn_output.weight
Q5_K
Q5_K
[6144, 3072]
blk.5.attn_q.weight
Q5_K
Q5_K
[3072, 6144]
blk.5.attn_q_norm.weight
F32
F32
[6144]
blk.5.attn_v.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.5.exp_probs_b.bias
F32
F32
[256]
blk.5.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.5.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.5.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.5.ffn_norm.weight
F32
F32
[3072]
blk.5.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.6
blk.6.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.6.attn_k_norm.weight
F32
F32
[1024]
blk.6.attn_norm.weight
F32
F32
[3072]
blk.6.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.6.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.6.attn_q_norm.weight
F32
F32
[6144]
blk.6.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.6.exp_probs_b.bias
F32
F32
[256]
blk.6.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.6.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.6.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.6.ffn_norm.weight
F32
F32
[3072]
blk.6.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.7
blk.7.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.7.attn_k_norm.weight
F32
F32
[1024]
blk.7.attn_norm.weight
F32
F32
[3072]
blk.7.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.7.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.7.attn_q_norm.weight
F32
F32
[6144]
blk.7.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.7.exp_probs_b.bias
F32
F32
[256]
blk.7.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.7.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.7.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.7.ffn_norm.weight
F32
F32
[3072]
blk.7.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.8
blk.8.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.8.attn_k_norm.weight
F32
F32
[1024]
blk.8.attn_norm.weight
F32
F32
[3072]
blk.8.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.8.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.8.attn_q_norm.weight
F32
F32
[6144]
blk.8.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.8.exp_probs_b.bias
F32
F32
[256]
blk.8.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.8.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.8.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.8.ffn_norm.weight
F32
F32
[3072]
blk.8.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.9
blk.9.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.9.attn_k_norm.weight
F32
F32
[1024]
blk.9.attn_norm.weight
F32
F32
[3072]
blk.9.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.9.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.9.attn_q_norm.weight
F32
F32
[6144]
blk.9.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.9.exp_probs_b.bias
F32
F32
[256]
blk.9.ffn_down_exps.weight
IQ3_S
IQ3_S
[1536, 3072, 256]
blk.9.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.9.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.9.ffn_norm.weight
F32
F32
[3072]
blk.9.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.10
blk.10.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.10.attn_k_norm.weight
F32
F32
[1024]
blk.10.attn_norm.weight
F32
F32
[3072]
blk.10.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.10.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.10.attn_q_norm.weight
F32
F32
[6144]
blk.10.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.10.exp_probs_b.bias
F32
F32
[256]
blk.10.ffn_down_exps.weight
IQ3_S
IQ3_S
[1536, 3072, 256]
blk.10.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.10.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.10.ffn_norm.weight
F32
F32
[3072]
blk.10.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.11
blk.11.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.11.attn_k_norm.weight
F32
F32
[1024]
blk.11.attn_norm.weight
F32
F32
[3072]
blk.11.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.11.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.11.attn_q_norm.weight
F32
F32
[6144]
blk.11.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.11.exp_probs_b.bias
F32
F32
[256]
blk.11.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.11.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.11.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.11.ffn_norm.weight
F32
F32
[3072]
blk.11.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.12
blk.12.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.12.attn_k_norm.weight
F32
F32
[1024]
blk.12.attn_norm.weight
F32
F32
[3072]
blk.12.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.12.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.12.attn_q_norm.weight
F32
F32
[6144]
blk.12.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.12.exp_probs_b.bias
F32
F32
[256]
blk.12.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.12.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.12.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.12.ffn_norm.weight
F32
F32
[3072]
blk.12.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.13
blk.13.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.13.attn_k_norm.weight
F32
F32
[1024]
blk.13.attn_norm.weight
F32
F32
[3072]
blk.13.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.13.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.13.attn_q_norm.weight
F32
F32
[6144]
blk.13.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.13.exp_probs_b.bias
F32
F32
[256]
blk.13.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.13.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.13.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.13.ffn_norm.weight
F32
F32
[3072]
blk.13.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.14
blk.14.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.14.attn_k_norm.weight
F32
F32
[1024]
blk.14.attn_norm.weight
F32
F32
[3072]
blk.14.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.14.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.14.attn_q_norm.weight
F32
F32
[6144]
blk.14.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.14.exp_probs_b.bias
F32
F32
[256]
blk.14.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.14.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.14.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.14.ffn_norm.weight
F32
F32
[3072]
blk.14.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.15
blk.15.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.15.attn_k_norm.weight
F32
F32
[1024]
blk.15.attn_norm.weight
F32
F32
[3072]
blk.15.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.15.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.15.attn_q_norm.weight
F32
F32
[6144]
blk.15.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.15.exp_probs_b.bias
F32
F32
[256]
blk.15.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.15.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.15.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.15.ffn_norm.weight
F32
F32
[3072]
blk.15.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.16
blk.16.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.16.attn_k_norm.weight
F32
F32
[1024]
blk.16.attn_norm.weight
F32
F32
[3072]
blk.16.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.16.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.16.attn_q_norm.weight
F32
F32
[6144]
blk.16.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.16.exp_probs_b.bias
F32
F32
[256]
blk.16.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.16.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.16.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.16.ffn_norm.weight
F32
F32
[3072]
blk.16.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.17
blk.17.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.17.attn_k_norm.weight
F32
F32
[1024]
blk.17.attn_norm.weight
F32
F32
[3072]
blk.17.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.17.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.17.attn_q_norm.weight
F32
F32
[6144]
blk.17.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.17.exp_probs_b.bias
F32
F32
[256]
blk.17.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.17.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.17.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.17.ffn_norm.weight
F32
F32
[3072]
blk.17.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.18
blk.18.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.18.attn_k_norm.weight
F32
F32
[1024]
blk.18.attn_norm.weight
F32
F32
[3072]
blk.18.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.18.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.18.attn_q_norm.weight
F32
F32
[6144]
blk.18.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.18.exp_probs_b.bias
F32
F32
[256]
blk.18.ffn_down_exps.weight
IQ3_S
IQ3_S
[1536, 3072, 256]
blk.18.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.18.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.18.ffn_norm.weight
F32
F32
[3072]
blk.18.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.19
blk.19.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.19.attn_k_norm.weight
F32
F32
[1024]
blk.19.attn_norm.weight
F32
F32
[3072]
blk.19.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.19.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.19.attn_q_norm.weight
F32
F32
[6144]
blk.19.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.19.exp_probs_b.bias
F32
F32
[256]
blk.19.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.19.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.19.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.19.ffn_norm.weight
F32
F32
[3072]
blk.19.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.20
blk.20.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.20.attn_k_norm.weight
F32
F32
[1024]
blk.20.attn_norm.weight
F32
F32
[3072]
blk.20.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.20.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.20.attn_q_norm.weight
F32
F32
[6144]
blk.20.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.20.exp_probs_b.bias
F32
F32
[256]
blk.20.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.20.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.20.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.20.ffn_norm.weight
F32
F32
[3072]
blk.20.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.21
blk.21.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.21.attn_k_norm.weight
F32
F32
[1024]
blk.21.attn_norm.weight
F32
F32
[3072]
blk.21.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.21.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.21.attn_q_norm.weight
F32
F32
[6144]
blk.21.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.21.exp_probs_b.bias
F32
F32
[256]
blk.21.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.21.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.21.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.21.ffn_norm.weight
F32
F32
[3072]
blk.21.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.22
blk.22.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.22.attn_k_norm.weight
F32
F32
[1024]
blk.22.attn_norm.weight
F32
F32
[3072]
blk.22.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.22.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.22.attn_q_norm.weight
F32
F32
[6144]
blk.22.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.22.exp_probs_b.bias
F32
F32
[256]
blk.22.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.22.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.22.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.22.ffn_norm.weight
F32
F32
[3072]
blk.22.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.23
blk.23.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.23.attn_k_norm.weight
F32
F32
[1024]
blk.23.attn_norm.weight
F32
F32
[3072]
blk.23.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.23.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.23.attn_q_norm.weight
F32
F32
[6144]
blk.23.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.23.exp_probs_b.bias
F32
F32
[256]
blk.23.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.23.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.23.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.23.ffn_norm.weight
F32
F32
[3072]
blk.23.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.24
blk.24.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.24.attn_k_norm.weight
F32
F32
[1024]
blk.24.attn_norm.weight
F32
F32
[3072]
blk.24.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.24.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.24.attn_q_norm.weight
F32
F32
[6144]
blk.24.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.24.exp_probs_b.bias
F32
F32
[256]
blk.24.ffn_down_exps.weight
IQ3_S
IQ3_S
[1536, 3072, 256]
blk.24.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.24.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.24.ffn_norm.weight
F32
F32
[3072]
blk.24.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.25
blk.25.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.25.attn_k_norm.weight
F32
F32
[1024]
blk.25.attn_norm.weight
F32
F32
[3072]
blk.25.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.25.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.25.attn_q_norm.weight
F32
F32
[6144]
blk.25.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.25.exp_probs_b.bias
F32
F32
[256]
blk.25.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.25.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.25.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.25.ffn_norm.weight
F32
F32
[3072]
blk.25.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.26
blk.26.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.26.attn_k_norm.weight
F32
F32
[1024]
blk.26.attn_norm.weight
F32
F32
[3072]
blk.26.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.26.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.26.attn_q_norm.weight
F32
F32
[6144]
blk.26.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.26.exp_probs_b.bias
F32
F32
[256]
blk.26.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.26.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.26.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.26.ffn_norm.weight
F32
F32
[3072]
blk.26.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.27
blk.27.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.27.attn_k_norm.weight
F32
F32
[1024]
blk.27.attn_norm.weight
F32
F32
[3072]
blk.27.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.27.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.27.attn_q_norm.weight
F32
F32
[6144]
blk.27.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.27.exp_probs_b.bias
F32
F32
[256]
blk.27.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.27.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.27.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.27.ffn_norm.weight
F32
F32
[3072]
blk.27.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.28
blk.28.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.28.attn_k_norm.weight
F32
F32
[1024]
blk.28.attn_norm.weight
F32
F32
[3072]
blk.28.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.28.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.28.attn_q_norm.weight
F32
F32
[6144]
blk.28.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.28.exp_probs_b.bias
F32
F32
[256]
blk.28.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.28.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.28.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.28.ffn_norm.weight
F32
F32
[3072]
blk.28.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.29
blk.29.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.29.attn_k_norm.weight
F32
F32
[1024]
blk.29.attn_norm.weight
F32
F32
[3072]
blk.29.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.29.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.29.attn_q_norm.weight
F32
F32
[6144]
blk.29.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.29.exp_probs_b.bias
F32
F32
[256]
blk.29.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.29.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.29.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.29.ffn_norm.weight
F32
F32
[3072]
blk.29.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.30
blk.30.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.30.attn_k_norm.weight
F32
F32
[1024]
blk.30.attn_norm.weight
F32
F32
[3072]
blk.30.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.30.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.30.attn_q_norm.weight
F32
F32
[6144]
blk.30.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.30.exp_probs_b.bias
F32
F32
[256]
blk.30.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.30.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.30.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.30.ffn_norm.weight
F32
F32
[3072]
blk.30.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.31
blk.31.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.31.attn_k_norm.weight
F32
F32
[1024]
blk.31.attn_norm.weight
F32
F32
[3072]
blk.31.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.31.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.31.attn_q_norm.weight
F32
F32
[6144]
blk.31.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.31.exp_probs_b.bias
F32
F32
[256]
blk.31.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.31.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.31.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.31.ffn_norm.weight
F32
F32
[3072]
blk.31.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.32
blk.32.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.32.attn_k_norm.weight
F32
F32
[1024]
blk.32.attn_norm.weight
F32
F32
[3072]
blk.32.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.32.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.32.attn_q_norm.weight
F32
F32
[6144]
blk.32.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.32.exp_probs_b.bias
F32
F32
[256]
blk.32.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.32.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.32.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.32.ffn_norm.weight
F32
F32
[3072]
blk.32.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.33
blk.33.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.33.attn_k_norm.weight
F32
F32
[1024]
blk.33.attn_norm.weight
F32
F32
[3072]
blk.33.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.33.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.33.attn_q_norm.weight
F32
F32
[6144]
blk.33.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.33.exp_probs_b.bias
F32
F32
[256]
blk.33.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.33.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.33.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.33.ffn_norm.weight
F32
F32
[3072]
blk.33.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.34
blk.34.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.34.attn_k_norm.weight
F32
F32
[1024]
blk.34.attn_norm.weight
F32
F32
[3072]
blk.34.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.34.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.34.attn_q_norm.weight
F32
F32
[6144]
blk.34.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.34.exp_probs_b.bias
F32
F32
[256]
blk.34.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.34.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.34.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.34.ffn_norm.weight
F32
F32
[3072]
blk.34.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.35
blk.35.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.35.attn_k_norm.weight
F32
F32
[1024]
blk.35.attn_norm.weight
F32
F32
[3072]
blk.35.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.35.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.35.attn_q_norm.weight
F32
F32
[6144]
blk.35.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.35.exp_probs_b.bias
F32
F32
[256]
blk.35.ffn_down_exps.weight
IQ2_XXS
IQ2_XXS
[1536, 3072, 256]
blk.35.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.35.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.35.ffn_norm.weight
F32
F32
[3072]
blk.35.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.36
blk.36.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.36.attn_k_norm.weight
F32
F32
[1024]
blk.36.attn_norm.weight
F32
F32
[3072]
blk.36.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.36.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.36.attn_q_norm.weight
F32
F32
[6144]
blk.36.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.36.exp_probs_b.bias
F32
F32
[256]
blk.36.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.36.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.36.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.36.ffn_norm.weight
F32
F32
[3072]
blk.36.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.37
blk.37.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.37.attn_k_norm.weight
F32
F32
[1024]
blk.37.attn_norm.weight
F32
F32
[3072]
blk.37.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.37.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.37.attn_q_norm.weight
F32
F32
[6144]
blk.37.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.37.exp_probs_b.bias
F32
F32
[256]
blk.37.ffn_down_exps.weight
IQ2_XXS
IQ2_XXS
[1536, 3072, 256]
blk.37.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.37.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.37.ffn_norm.weight
F32
F32
[3072]
blk.37.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.38
blk.38.attn_k.weight
Q5_K
Q5_K
[3072, 1024]
blk.38.attn_k_norm.weight
F32
F32
[1024]
blk.38.attn_norm.weight
F32
F32
[3072]
blk.38.attn_output.weight
Q5_K
Q5_K
[6144, 3072]
blk.38.attn_q.weight
Q5_K
Q5_K
[3072, 6144]
blk.38.attn_q_norm.weight
F32
F32
[6144]
blk.38.attn_v.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.38.exp_probs_b.bias
F32
F32
[256]
blk.38.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.38.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.38.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.38.ffn_norm.weight
F32
F32
[3072]
blk.38.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.39
blk.39.attn_k.weight
Q5_K
Q5_K
[3072, 1024]
blk.39.attn_k_norm.weight
F32
F32
[1024]
blk.39.attn_norm.weight
F32
F32
[3072]
blk.39.attn_output.weight
Q5_K
Q5_K
[6144, 3072]
blk.39.attn_q.weight
Q5_K
Q5_K
[3072, 6144]
blk.39.attn_q_norm.weight
F32
F32
[6144]
blk.39.attn_v.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.39.exp_probs_b.bias
F32
F32
[256]
blk.39.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.39.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.39.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.39.ffn_norm.weight
F32
F32
[3072]
blk.39.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.40
blk.40.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.40.attn_k_norm.weight
F32
F32
[1024]
blk.40.attn_norm.weight
F32
F32
[3072]
blk.40.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.40.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.40.attn_q_norm.weight
F32
F32
[6144]
blk.40.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.40.exp_probs_b.bias
F32
F32
[256]
blk.40.ffn_down_exps.weight
IQ2_XXS
IQ2_XXS
[1536, 3072, 256]
blk.40.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.40.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.40.ffn_norm.weight
F32
F32
[3072]
blk.40.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.41
blk.41.attn_k.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.41.attn_k_norm.weight
F32
F32
[1024]
blk.41.attn_norm.weight
F32
F32
[3072]
blk.41.attn_output.weight
IQ4_XS
IQ4_XS
[6144, 3072]
blk.41.attn_q.weight
IQ4_XS
IQ4_XS
[3072, 6144]
blk.41.attn_q_norm.weight
F32
F32
[6144]
blk.41.attn_v.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.41.exp_probs_b.bias
F32
F32
[256]
blk.41.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[1536, 3072, 256]
blk.41.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.41.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.41.ffn_norm.weight
F32
F32
[3072]
blk.41.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.42
blk.42.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.42.attn_k_norm.weight
F32
F32
[1024]
blk.42.attn_norm.weight
F32
F32
[3072]
blk.42.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.42.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.42.attn_q_norm.weight
F32
F32
[6144]
blk.42.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.42.exp_probs_b.bias
F32
F32
[256]
blk.42.ffn_down_exps.weight
IQ2_XXS
IQ2_XXS
[1536, 3072, 256]
blk.42.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.42.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.42.ffn_norm.weight
F32
F32
[3072]
blk.42.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.43
blk.43.attn_k.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.43.attn_k_norm.weight
F32
F32
[1024]
blk.43.attn_norm.weight
F32
F32
[3072]
blk.43.attn_output.weight
IQ4_XS
IQ4_XS
[6144, 3072]
blk.43.attn_q.weight
IQ4_XS
IQ4_XS
[3072, 6144]
blk.43.attn_q_norm.weight
F32
F32
[6144]
blk.43.attn_v.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.43.exp_probs_b.bias
F32
F32
[256]
blk.43.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[1536, 3072, 256]
blk.43.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.43.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.43.ffn_norm.weight
F32
F32
[3072]
blk.43.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.44
blk.44.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.44.attn_k_norm.weight
F32
F32
[1024]
blk.44.attn_norm.weight
F32
F32
[3072]
blk.44.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.44.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.44.attn_q_norm.weight
F32
F32
[6144]
blk.44.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.44.exp_probs_b.bias
F32
F32
[256]
blk.44.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.44.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.44.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.44.ffn_norm.weight
F32
F32
[3072]
blk.44.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.45
blk.45.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.45.attn_k_norm.weight
F32
F32
[1024]
blk.45.attn_norm.weight
F32
F32
[3072]
blk.45.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.45.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.45.attn_q_norm.weight
F32
F32
[6144]
blk.45.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.45.exp_probs_b.bias
F32
F32
[256]
blk.45.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.45.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.45.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.45.ffn_norm.weight
F32
F32
[3072]
blk.45.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.46
blk.46.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.46.attn_k_norm.weight
F32
F32
[1024]
blk.46.attn_norm.weight
F32
F32
[3072]
blk.46.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.46.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.46.attn_q_norm.weight
F32
F32
[6144]
blk.46.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.46.exp_probs_b.bias
F32
F32
[256]
blk.46.ffn_down_exps.weight
IQ2_XXS
IQ2_XXS
[1536, 3072, 256]
blk.46.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.46.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.46.ffn_norm.weight
F32
F32
[3072]
blk.46.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.47
blk.47.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.47.attn_k_norm.weight
F32
F32
[1024]
blk.47.attn_norm.weight
F32
F32
[3072]
blk.47.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.47.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.47.attn_q_norm.weight
F32
F32
[6144]
blk.47.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.47.exp_probs_b.bias
F32
F32
[256]
blk.47.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.47.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.47.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.47.ffn_norm.weight
F32
F32
[3072]
blk.47.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.48
blk.48.attn_k.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.48.attn_k_norm.weight
F32
F32
[1024]
blk.48.attn_norm.weight
F32
F32
[3072]
blk.48.attn_output.weight
IQ4_XS
IQ4_XS
[6144, 3072]
blk.48.attn_q.weight
IQ4_XS
IQ4_XS
[3072, 6144]
blk.48.attn_q_norm.weight
F32
F32
[6144]
blk.48.attn_v.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.48.exp_probs_b.bias
F32
F32
[256]
blk.48.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[1536, 3072, 256]
blk.48.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.48.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.48.ffn_norm.weight
F32
F32
[3072]
blk.48.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.49
blk.49.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.49.attn_k_norm.weight
F32
F32
[1024]
blk.49.attn_norm.weight
F32
F32
[3072]
blk.49.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.49.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.49.attn_q_norm.weight
F32
F32
[6144]
blk.49.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.49.exp_probs_b.bias
F32
F32
[256]
blk.49.ffn_down_exps.weight
IQ2_XXS
IQ2_XXS
[1536, 3072, 256]
blk.49.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.49.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.49.ffn_norm.weight
F32
F32
[3072]
blk.49.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.50
blk.50.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.50.attn_k_norm.weight
F32
F32
[1024]
blk.50.attn_norm.weight
F32
F32
[3072]
blk.50.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.50.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.50.attn_q_norm.weight
F32
F32
[6144]
blk.50.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.50.exp_probs_b.bias
F32
F32
[256]
blk.50.ffn_down_exps.weight
IQ3_S
IQ3_S
[1536, 3072, 256]
blk.50.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.50.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.50.ffn_norm.weight
F32
F32
[3072]
blk.50.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.51
blk.51.attn_k.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.51.attn_k_norm.weight
F32
F32
[1024]
blk.51.attn_norm.weight
F32
F32
[3072]
blk.51.attn_output.weight
IQ4_XS
IQ4_XS
[6144, 3072]
blk.51.attn_q.weight
IQ4_XS
IQ4_XS
[3072, 6144]
blk.51.attn_q_norm.weight
F32
F32
[6144]
blk.51.attn_v.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.51.exp_probs_b.bias
F32
F32
[256]
blk.51.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[1536, 3072, 256]
blk.51.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.51.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.51.ffn_norm.weight
F32
F32
[3072]
blk.51.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.52
blk.52.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.52.attn_k_norm.weight
F32
F32
[1024]
blk.52.attn_norm.weight
F32
F32
[3072]
blk.52.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.52.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.52.attn_q_norm.weight
F32
F32
[6144]
blk.52.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.52.exp_probs_b.bias
F32
F32
[256]
blk.52.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.52.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.52.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.52.ffn_norm.weight
F32
F32
[3072]
blk.52.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.53
blk.53.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.53.attn_k_norm.weight
F32
F32
[1024]
blk.53.attn_norm.weight
F32
F32
[3072]
blk.53.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.53.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.53.attn_q_norm.weight
F32
F32
[6144]
blk.53.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.53.exp_probs_b.bias
F32
F32
[256]
blk.53.ffn_down_exps.weight
IQ3_S
IQ3_S
[1536, 3072, 256]
blk.53.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.53.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.53.ffn_norm.weight
F32
F32
[3072]
blk.53.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.54
blk.54.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.54.attn_k_norm.weight
F32
F32
[1024]
blk.54.attn_norm.weight
F32
F32
[3072]
blk.54.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.54.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.54.attn_q_norm.weight
F32
F32
[6144]
blk.54.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.54.exp_probs_b.bias
F32
F32
[256]
blk.54.ffn_down_exps.weight
IQ2_XXS
IQ2_XXS
[1536, 3072, 256]
blk.54.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.54.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.54.ffn_norm.weight
F32
F32
[3072]
blk.54.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.55
blk.55.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.55.attn_k_norm.weight
F32
F32
[1024]
blk.55.attn_norm.weight
F32
F32
[3072]
blk.55.attn_output.weight
Q6_K
Q6_K
[6144, 3072]
blk.55.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.55.attn_q_norm.weight
F32
F32
[6144]
blk.55.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.55.exp_probs_b.bias
F32
F32
[256]
blk.55.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.55.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.55.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.55.ffn_norm.weight
F32
F32
[3072]
blk.55.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.56
blk.56.attn_k.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.56.attn_k_norm.weight
F32
F32
[1024]
blk.56.attn_norm.weight
F32
F32
[3072]
blk.56.attn_output.weight
IQ3_S
IQ3_S
[6144, 3072]
blk.56.attn_q.weight
IQ3_XXS
IQ3_XXS
[3072, 6144]
blk.56.attn_q_norm.weight
F32
F32
[6144]
blk.56.attn_v.weight
IQ3_XXS
IQ3_XXS
[3072, 1024]
blk.56.exp_probs_b.bias
F32
F32
[256]
blk.56.ffn_down_exps.weight
IQ1_S
IQ1_S
[1536, 3072, 256]
blk.56.ffn_gate_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.56.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.56.ffn_norm.weight
F32
F32
[3072]
blk.56.ffn_up_exps.weight
IQ1_S
IQ1_S
[3072, 1536, 256]
blk.57
blk.57.attn_k.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.57.attn_k_norm.weight
F32
F32
[1024]
blk.57.attn_norm.weight
F32
F32
[3072]
blk.57.attn_output.weight
IQ4_XS
IQ4_XS
[6144, 3072]
blk.57.attn_q.weight
IQ4_XS
IQ4_XS
[3072, 6144]
blk.57.attn_q_norm.weight
F32
F32
[6144]
blk.57.attn_v.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.57.exp_probs_b.bias
F32
F32
[256]
blk.57.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[1536, 3072, 256]
blk.57.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.57.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.57.ffn_norm.weight
F32
F32
[3072]
blk.57.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.58
blk.58.attn_k.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.58.attn_k_norm.weight
F32
F32
[1024]
blk.58.attn_norm.weight
F32
F32
[3072]
blk.58.attn_output.weight
IQ4_XS
IQ4_XS
[6144, 3072]
blk.58.attn_q.weight
IQ4_XS
IQ4_XS
[3072, 6144]
blk.58.attn_q_norm.weight
F32
F32
[6144]
blk.58.attn_v.weight
IQ3_S
IQ3_S
[3072, 1024]
blk.58.exp_probs_b.bias
F32
F32
[256]
blk.58.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[1536, 3072, 256]
blk.58.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.58.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.58.ffn_norm.weight
F32
F32
[3072]
blk.58.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.59
blk.59.attn_k.weight
Q5_K
Q5_K
[3072, 1024]
blk.59.attn_k_norm.weight
F32
F32
[1024]
blk.59.attn_norm.weight
F32
F32
[3072]
blk.59.attn_output.weight
Q5_K
Q5_K
[6144, 3072]
blk.59.attn_q.weight
Q5_K
Q5_K
[3072, 6144]
blk.59.attn_q_norm.weight
F32
F32
[6144]
blk.59.attn_v.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.59.exp_probs_b.bias
F32
F32
[256]
blk.59.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.59.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.59.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.59.ffn_norm.weight
F32
F32
[3072]
blk.59.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.60
blk.60.attn_k.weight
Q5_K
Q5_K
[3072, 1024]
blk.60.attn_k_norm.weight
F32
F32
[1024]
blk.60.attn_norm.weight
F32
F32
[3072]
blk.60.attn_output.weight
Q5_K
Q5_K
[6144, 3072]
blk.60.attn_q.weight
Q5_K
Q5_K
[3072, 6144]
blk.60.attn_q_norm.weight
F32
F32
[6144]
blk.60.attn_v.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.60.exp_probs_b.bias
F32
F32
[256]
blk.60.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.60.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.60.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.60.ffn_norm.weight
F32
F32
[3072]
blk.60.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.61
blk.61.attn_k.weight
Q5_K
Q5_K
[3072, 1024]
blk.61.attn_k_norm.weight
F32
F32
[1024]
blk.61.attn_norm.weight
F32
F32
[3072]
blk.61.attn_output.weight
Q5_K
Q5_K
[6144, 3072]
blk.61.attn_q.weight
Q5_K
Q5_K
[3072, 6144]
blk.61.attn_q_norm.weight
F32
F32
[6144]
blk.61.attn_v.weight
IQ4_XS
IQ4_XS
[3072, 1024]
blk.61.exp_probs_b.bias
F32
F32
[256]
blk.61.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[1536, 3072, 256]
blk.61.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
blk.61.ffn_gate_inp.weight
F32
F32
[3072, 256]
blk.61.ffn_norm.weight
F32
F32
[3072]
blk.61.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[3072, 1536, 256]
output.weight
Q6_K
Q6_K
[3072, 200064]
output_norm.weight
F32
F32
[3072]