GitHub
185.7K
Download
Models
Docs
Pricing
Sign in
Download
Models
Docs
Pricing
Sign in
frob
/
deepseek-v4.1-flash
:552b-a8b-1017-ud-iq1_M
205
Downloads
Updated
5 days ago
1017
1017
Cancel
vision
552b
deepseek-v4.1-flash:552b-a8b-1017-ud-iq1_M
...
/
model
28e0edad226a ยท 41GB
Metadata
split.count
5
5
split.no
3
3
split.tensors.count
1043
1043
Tensor
Name
Type
Shape
blk.29
blk.29.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.29.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.29.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.29.ffn_norm.weight
F32
F32
[5120]
blk.29.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.29.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.29.hc_attn_base.weight
F32
F32
[24]
blk.29.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.29.hc_attn_scale.weight
F32
F32
[3]
blk.29.hc_ffn_base.weight
F32
F32
[24]
blk.29.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.29.hc_ffn_scale.weight
F32
F32
[3]
blk.30
blk.30.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.30.attn_kv_a_norm.weight
F32
F32
[512]
blk.30.attn_norm.weight
F32
F32
[5120]
blk.30.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.30.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.30.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.30.attn_q_a_norm.weight
F32
F32
[1280]
blk.30.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.30.attn_sinks.weight
F32
F32
[64]
blk.30.exp_probs_b.bias
F32
F32
[384]
blk.30.exp_probs_b_vl.bias
F32
F32
[384]
blk.30.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.30.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.30.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.30.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.30.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.30.ffn_norm.weight
F32
F32
[5120]
blk.30.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.30.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.30.hc_attn_base.weight
F32
F32
[24]
blk.30.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.30.hc_attn_scale.weight
F32
F32
[3]
blk.30.hc_ffn_base.weight
F32
F32
[24]
blk.30.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.30.hc_ffn_scale.weight
F32
F32
[3]
blk.31
blk.31.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.31.attn_kv_a_norm.weight
F32
F32
[512]
blk.31.attn_norm.weight
F32
F32
[5120]
blk.31.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.31.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.31.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.31.attn_q_a_norm.weight
F32
F32
[1280]
blk.31.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.31.attn_sinks.weight
F32
F32
[64]
blk.31.exp_probs_b.bias
F32
F32
[384]
blk.31.exp_probs_b_vl.bias
F32
F32
[384]
blk.31.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.31.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.31.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.31.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.31.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.31.ffn_norm.weight
F32
F32
[5120]
blk.31.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.31.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.31.hc_attn_base.weight
F32
F32
[24]
blk.31.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.31.hc_attn_scale.weight
F32
F32
[3]
blk.31.hc_ffn_base.weight
F32
F32
[24]
blk.31.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.31.hc_ffn_scale.weight
F32
F32
[3]
blk.32
blk.32.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.32.attn_kv_a_norm.weight
F32
F32
[512]
blk.32.attn_norm.weight
F32
F32
[5120]
blk.32.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.32.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.32.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.32.attn_q_a_norm.weight
F32
F32
[1280]
blk.32.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.32.attn_sinks.weight
F32
F32
[64]
blk.32.exp_probs_b.bias
F32
F32
[384]
blk.32.exp_probs_b_vl.bias
F32
F32
[384]
blk.32.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.32.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.32.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.32.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.32.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.32.ffn_norm.weight
F32
F32
[5120]
blk.32.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.32.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.32.hc_attn_base.weight
F32
F32
[24]
blk.32.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.32.hc_attn_scale.weight
F32
F32
[3]
blk.32.hc_ffn_base.weight
F32
F32
[24]
blk.32.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.32.hc_ffn_scale.weight
F32
F32
[3]
blk.32.indexer.attn_q_b.weight
BF16
BF16
[1280, 4096]
blk.32.indexer.proj.weight
BF16
BF16
[5120, 32]
blk.33
blk.33.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.33.attn_kv_a_norm.weight
F32
F32
[512]
blk.33.attn_norm.weight
F32
F32
[5120]
blk.33.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.33.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.33.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.33.attn_q_a_norm.weight
F32
F32
[1280]
blk.33.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.33.attn_sinks.weight
F32
F32
[64]
blk.33.exp_probs_b.bias
F32
F32
[384]
blk.33.exp_probs_b_vl.bias
F32
F32
[384]
blk.33.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.33.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.33.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.33.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.33.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.33.ffn_norm.weight
F32
F32
[5120]
blk.33.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.33.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.33.hc_attn_base.weight
F32
F32
[24]
blk.33.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.33.hc_attn_scale.weight
F32
F32
[3]
blk.33.hc_ffn_base.weight
F32
F32
[24]
blk.33.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.33.hc_ffn_scale.weight
F32
F32
[3]
blk.34
blk.34.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.34.attn_kv_a_norm.weight
F32
F32
[512]
blk.34.attn_norm.weight
F32
F32
[5120]
blk.34.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.34.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.34.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.34.attn_q_a_norm.weight
F32
F32
[1280]
blk.34.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.34.attn_sinks.weight
F32
F32
[64]
blk.34.exp_probs_b.bias
F32
F32
[384]
blk.34.exp_probs_b_vl.bias
F32
F32
[384]
blk.34.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.34.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.34.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.34.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.34.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.34.ffn_norm.weight
F32
F32
[5120]
blk.34.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.34.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.34.hc_attn_base.weight
F32
F32
[24]
blk.34.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.34.hc_attn_scale.weight
F32
F32
[3]
blk.34.hc_ffn_base.weight
F32
F32
[24]
blk.34.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.34.hc_ffn_scale.weight
F32
F32
[3]
blk.35
blk.35.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.35.attn_kv_a_norm.weight
F32
F32
[512]
blk.35.attn_norm.weight
F32
F32
[5120]
blk.35.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.35.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.35.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.35.attn_q_a_norm.weight
F32
F32
[1280]
blk.35.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.35.attn_sinks.weight
F32
F32
[64]
blk.35.exp_probs_b.bias
F32
F32
[384]
blk.35.exp_probs_b_vl.bias
F32
F32
[384]
blk.35.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.35.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.35.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.35.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.35.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.35.ffn_norm.weight
F32
F32
[5120]
blk.35.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.35.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.35.hc_attn_base.weight
F32
F32
[24]
blk.35.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.35.hc_attn_scale.weight
F32
F32
[3]
blk.35.hc_ffn_base.weight
F32
F32
[24]
blk.35.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.35.hc_ffn_scale.weight
F32
F32
[3]
blk.36
blk.36.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.36.attn_kv_a_norm.weight
F32
F32
[512]
blk.36.attn_norm.weight
F32
F32
[5120]
blk.36.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.36.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.36.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.36.attn_q_a_norm.weight
F32
F32
[1280]
blk.36.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.36.attn_sinks.weight
F32
F32
[64]
blk.36.exp_probs_b.bias
F32
F32
[384]
blk.36.exp_probs_b_vl.bias
F32
F32
[384]
blk.36.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.36.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.36.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.36.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.36.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.36.ffn_norm.weight
F32
F32
[5120]
blk.36.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.36.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.36.hc_attn_base.weight
F32
F32
[24]
blk.36.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.36.hc_attn_scale.weight
F32
F32
[3]
blk.36.hc_ffn_base.weight
F32
F32
[24]
blk.36.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.36.hc_ffn_scale.weight
F32
F32
[3]
blk.36.indexer.attn_q_b.weight
BF16
BF16
[1280, 4096]
blk.36.indexer.proj.weight
BF16
BF16
[5120, 32]
blk.37
blk.37.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.37.attn_kv_a_norm.weight
F32
F32
[512]
blk.37.attn_norm.weight
F32
F32
[5120]
blk.37.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.37.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.37.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.37.attn_q_a_norm.weight
F32
F32
[1280]
blk.37.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.37.attn_sinks.weight
F32
F32
[64]
blk.37.exp_probs_b.bias
F32
F32
[384]
blk.37.exp_probs_b_vl.bias
F32
F32
[384]
blk.37.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.37.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.37.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.37.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.37.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.37.ffn_norm.weight
F32
F32
[5120]
blk.37.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[5120, 2304, 384]
blk.37.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.37.hc_attn_base.weight
F32
F32
[24]
blk.37.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.37.hc_attn_scale.weight
F32
F32
[3]
blk.37.hc_ffn_base.weight
F32
F32
[24]
blk.37.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.37.hc_ffn_scale.weight
F32
F32
[3]
blk.38
blk.38.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.38.attn_kv_a_norm.weight
F32
F32
[512]
blk.38.attn_norm.weight
F32
F32
[5120]
blk.38.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.38.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.38.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.38.attn_q_a_norm.weight
F32
F32
[1280]
blk.38.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.38.attn_sinks.weight
F32
F32
[64]
blk.38.exp_probs_b.bias
F32
F32
[384]
blk.38.exp_probs_b_vl.bias
F32
F32
[384]
blk.38.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.38.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.38.ffn_gate_exps.weight
IQ2_XS
IQ2_XS
[5120, 2304, 384]
blk.38.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.38.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.38.ffn_norm.weight
F32
F32
[5120]
blk.38.ffn_up_exps.weight
IQ2_XS
IQ2_XS
[5120, 2304, 384]
blk.38.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.38.hc_attn_base.weight
F32
F32
[24]
blk.38.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.38.hc_attn_scale.weight
F32
F32
[3]
blk.38.hc_ffn_base.weight
F32
F32
[24]
blk.38.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.38.hc_ffn_scale.weight
F32
F32
[3]
blk.39
blk.39.attn_kv.weight
Q6_K
Q6_K
[5120, 512]
blk.39.attn_kv_a_norm.weight
F32
F32
[512]
blk.39.attn_norm.weight
F32
F32
[5120]
blk.39.attn_output_a.weight
Q6_K
Q6_K
[4096, 8192]
blk.39.attn_output_b.weight
Q6_K
Q6_K
[8192, 5120]
blk.39.attn_q_a.weight
Q6_K
Q6_K
[5120, 1280]
blk.39.attn_q_a_norm.weight
F32
F32
[1280]
blk.39.attn_q_b.weight
Q6_K
Q6_K
[1280, 32768]
blk.39.attn_sinks.weight
F32
F32
[64]
blk.39.exp_probs_b.bias
F32
F32
[384]
blk.39.exp_probs_b_vl.bias
F32
F32
[384]
blk.39.ffn_down_exps.weight
IQ2_XS
IQ2_XS
[2304, 5120, 384]
blk.39.ffn_down_shexp.weight
Q5_K
Q5_K
[2304, 5120]
blk.39.ffn_gate_exps.weight
IQ2_XS
IQ2_XS
[5120, 2304, 384]
blk.39.ffn_gate_inp.weight
F32
F32
[5120, 384]
blk.39.ffn_gate_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.39.ffn_norm.weight
F32
F32
[5120]
blk.39.ffn_up_exps.weight
IQ2_XS
IQ2_XS
[5120, 2304, 384]
blk.39.ffn_up_shexp.weight
Q5_K
Q5_K
[5120, 2304]
blk.39.hc_attn_base.weight
F32
F32
[24]
blk.39.hc_attn_fn.weight
F32
F32
[20480, 24]
blk.39.hc_attn_scale.weight
F32
F32
[3]
blk.39.hc_ffn_base.weight
F32
F32
[24]
blk.39.hc_ffn_fn.weight
F32
F32
[20480, 24]
blk.39.hc_ffn_scale.weight
F32
F32
[3]