Models
GitHub
Discord
Docs
Cloud
Sign in
Download
Models
Download
GitHub
Discord
Docs
Cloud
Sign in
sam860
/
falcon-h1
:1.5b-deep-Q4_0
13
Downloads
Updated
1 month ago
Cancel
1.5b
falcon-h1:1.5b-deep-Q4_0
...
/
model
38d283df6414 · 904MB
Metadata
general.architecture
falcon-h1
falcon-h1
general.file_type
Q4_0
Q4_0
falcon-h1.attention.head_count
6
6
falcon-h1.attention.head_count_kv
2
2
falcon-h1.attention.key_length
128
128
falcon-h1.attention.layer_norm_rms_epsilon
1e-05
1e-05
falcon-h1.attention.value_length
128
128
falcon-h1.block_count
66
66
falcon-h1.context_length
131072
131072
falcon-h1.embedding_length
1280
1280
falcon-h1.feed_forward_length
3072
3072
falcon-h1.rope.freq_base
1e+11
1e+11
falcon-h1.ssm.conv_kernel
4
4
falcon-h1.ssm.group_count
1
1
falcon-h1.ssm.inner_size
1536
1536
falcon-h1.ssm.state_size
256
256
falcon-h1.ssm.time_step_rank
24
24
falcon-h1.vocab_size
65537
65537
tokenizer.ggml.bos_token_id
17
17
tokenizer.ggml.eos_token_id
228
228
tokenizer.ggml.merges
[Ġ Ġ, ĠĠ ĠĠ, Ġ t, i n, e r, ...]
[Ġ Ġ, ĠĠ ĠĠ, Ġ t, i n, e r, ...]
tokenizer.ggml.model
gpt2
gpt2
tokenizer.ggml.padding_token_id
65536
65536
tokenizer.ggml.pre
falcon-h1
falcon-h1
tokenizer.ggml.token_type
[3, 3, 3, 3, 3, ...]
[3, 3, 3, 3, 3, ...]
tokenizer.ggml.tokens
[<|pad|>, >>ABSTRACT<<, >>INTRODUCTION<<, >>SUMMARY<<, >>COMMENT<<, ...]
[<|pad|>, >>ABSTRACT<<, >>INTRODUCTION<<, >>SUMMARY<<, >>COMMENT<<, ...]
quantize.imatrix.chunks_count
582
582
quantize.imatrix.dataset
unsloth_calibration_Falcon-H1-1.5B-Deep-Instruct.txt
unsloth_calibration_Falcon-H1-1.5B-Deep-Instruct.txt
quantize.imatrix.entries_count
594
594
quantize.imatrix.file
Falcon-H1-1.5B-Deep-Instruct-GGUF/imatrix_unsloth.dat
Falcon-H1-1.5B-Deep-Instruct-GGUF/imatrix_unsloth.dat
Tensor
Name
Type
Shape
token_embd.weight
Q4_0
Q4_0
[1280, 65537]
blk.0
blk.0.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.0.attn_norm.weight
F32
F32
[1280]
blk.0.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.0.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.0.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.0.ffn_down.weight
Q4_1
Q4_1
[3072, 1280]
blk.0.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.0.ffn_norm
F32
F32
[1280]
blk.0.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.0.ssm_a
F32
F32
[1, 24]
blk.0.ssm_conv1d.bias
F32
F32
[2048]
blk.0.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.0.ssm_d
F32
F32
[1, 24]
blk.0.ssm_dt.bias
F32
F32
[24]
blk.0.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.0.ssm_norm.weight
F32
F32
[1536]
blk.0.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.1
blk.1.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.1.attn_norm.weight
F32
F32
[1280]
blk.1.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.1.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.1.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.1.ffn_down.weight
Q4_1
Q4_1
[3072, 1280]
blk.1.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.1.ffn_norm
F32
F32
[1280]
blk.1.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.1.ssm_a
F32
F32
[1, 24]
blk.1.ssm_conv1d.bias
F32
F32
[2048]
blk.1.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.1.ssm_d
F32
F32
[1, 24]
blk.1.ssm_dt.bias
F32
F32
[24]
blk.1.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.1.ssm_norm.weight
F32
F32
[1536]
blk.1.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.2
blk.2.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.2.attn_norm.weight
F32
F32
[1280]
blk.2.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.2.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.2.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.2.ffn_down.weight
Q4_1
Q4_1
[3072, 1280]
blk.2.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.2.ffn_norm
F32
F32
[1280]
blk.2.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.2.ssm_a
F32
F32
[1, 24]
blk.2.ssm_conv1d.bias
F32
F32
[2048]
blk.2.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.2.ssm_d
F32
F32
[1, 24]
blk.2.ssm_dt.bias
F32
F32
[24]
blk.2.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.2.ssm_norm.weight
F32
F32
[1536]
blk.2.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.3
blk.3.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.3.attn_norm.weight
F32
F32
[1280]
blk.3.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.3.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.3.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.3.ffn_down.weight
Q4_1
Q4_1
[3072, 1280]
blk.3.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.3.ffn_norm
F32
F32
[1280]
blk.3.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.3.ssm_a
F32
F32
[1, 24]
blk.3.ssm_conv1d.bias
F32
F32
[2048]
blk.3.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.3.ssm_d
F32
F32
[1, 24]
blk.3.ssm_dt.bias
F32
F32
[24]
blk.3.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.3.ssm_norm.weight
F32
F32
[1536]
blk.3.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.4
blk.4.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.4.attn_norm.weight
F32
F32
[1280]
blk.4.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.4.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.4.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.4.ffn_down.weight
Q4_1
Q4_1
[3072, 1280]
blk.4.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.4.ffn_norm
F32
F32
[1280]
blk.4.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.4.ssm_a
F32
F32
[1, 24]
blk.4.ssm_conv1d.bias
F32
F32
[2048]
blk.4.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.4.ssm_d
F32
F32
[1, 24]
blk.4.ssm_dt.bias
F32
F32
[24]
blk.4.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.4.ssm_norm.weight
F32
F32
[1536]
blk.4.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.5
blk.5.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.5.attn_norm.weight
F32
F32
[1280]
blk.5.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.5.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.5.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.5.ffn_down.weight
Q4_1
Q4_1
[3072, 1280]
blk.5.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.5.ffn_norm
F32
F32
[1280]
blk.5.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.5.ssm_a
F32
F32
[1, 24]
blk.5.ssm_conv1d.bias
F32
F32
[2048]
blk.5.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.5.ssm_d
F32
F32
[1, 24]
blk.5.ssm_dt.bias
F32
F32
[24]
blk.5.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.5.ssm_norm.weight
F32
F32
[1536]
blk.5.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.6
blk.6.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.6.attn_norm.weight
F32
F32
[1280]
blk.6.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.6.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.6.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.6.ffn_down.weight
Q4_1
Q4_1
[3072, 1280]
blk.6.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.6.ffn_norm
F32
F32
[1280]
blk.6.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.6.ssm_a
F32
F32
[1, 24]
blk.6.ssm_conv1d.bias
F32
F32
[2048]
blk.6.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.6.ssm_d
F32
F32
[1, 24]
blk.6.ssm_dt.bias
F32
F32
[24]
blk.6.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.6.ssm_norm.weight
F32
F32
[1536]
blk.6.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.7
blk.7.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.7.attn_norm.weight
F32
F32
[1280]
blk.7.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.7.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.7.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.7.ffn_down.weight
Q4_1
Q4_1
[3072, 1280]
blk.7.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.7.ffn_norm
F32
F32
[1280]
blk.7.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.7.ssm_a
F32
F32
[1, 24]
blk.7.ssm_conv1d.bias
F32
F32
[2048]
blk.7.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.7.ssm_d
F32
F32
[1, 24]
blk.7.ssm_dt.bias
F32
F32
[24]
blk.7.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.7.ssm_norm.weight
F32
F32
[1536]
blk.7.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.8
blk.8.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.8.attn_norm.weight
F32
F32
[1280]
blk.8.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.8.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.8.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.8.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.8.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.8.ffn_norm
F32
F32
[1280]
blk.8.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.8.ssm_a
F32
F32
[1, 24]
blk.8.ssm_conv1d.bias
F32
F32
[2048]
blk.8.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.8.ssm_d
F32
F32
[1, 24]
blk.8.ssm_dt.bias
F32
F32
[24]
blk.8.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.8.ssm_norm.weight
F32
F32
[1536]
blk.8.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.9
blk.9.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.9.attn_norm.weight
F32
F32
[1280]
blk.9.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.9.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.9.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.9.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.9.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.9.ffn_norm
F32
F32
[1280]
blk.9.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.9.ssm_a
F32
F32
[1, 24]
blk.9.ssm_conv1d.bias
F32
F32
[2048]
blk.9.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.9.ssm_d
F32
F32
[1, 24]
blk.9.ssm_dt.bias
F32
F32
[24]
blk.9.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.9.ssm_norm.weight
F32
F32
[1536]
blk.9.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.10
blk.10.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.10.attn_norm.weight
F32
F32
[1280]
blk.10.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.10.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.10.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.10.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.10.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.10.ffn_norm
F32
F32
[1280]
blk.10.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.10.ssm_a
F32
F32
[1, 24]
blk.10.ssm_conv1d.bias
F32
F32
[2048]
blk.10.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.10.ssm_d
F32
F32
[1, 24]
blk.10.ssm_dt.bias
F32
F32
[24]
blk.10.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.10.ssm_norm.weight
F32
F32
[1536]
blk.10.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.11
blk.11.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.11.attn_norm.weight
F32
F32
[1280]
blk.11.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.11.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.11.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.11.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.11.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.11.ffn_norm
F32
F32
[1280]
blk.11.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.11.ssm_a
F32
F32
[1, 24]
blk.11.ssm_conv1d.bias
F32
F32
[2048]
blk.11.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.11.ssm_d
F32
F32
[1, 24]
blk.11.ssm_dt.bias
F32
F32
[24]
blk.11.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.11.ssm_norm.weight
F32
F32
[1536]
blk.11.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.12
blk.12.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.12.attn_norm.weight
F32
F32
[1280]
blk.12.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.12.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.12.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.12.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.12.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.12.ffn_norm
F32
F32
[1280]
blk.12.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.12.ssm_a
F32
F32
[1, 24]
blk.12.ssm_conv1d.bias
F32
F32
[2048]
blk.12.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.12.ssm_d
F32
F32
[1, 24]
blk.12.ssm_dt.bias
F32
F32
[24]
blk.12.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.12.ssm_norm.weight
F32
F32
[1536]
blk.12.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.13
blk.13.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.13.attn_norm.weight
F32
F32
[1280]
blk.13.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.13.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.13.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.13.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.13.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.13.ffn_norm
F32
F32
[1280]
blk.13.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.13.ssm_a
F32
F32
[1, 24]
blk.13.ssm_conv1d.bias
F32
F32
[2048]
blk.13.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.13.ssm_d
F32
F32
[1, 24]
blk.13.ssm_dt.bias
F32
F32
[24]
blk.13.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.13.ssm_norm.weight
F32
F32
[1536]
blk.13.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.14
blk.14.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.14.attn_norm.weight
F32
F32
[1280]
blk.14.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.14.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.14.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.14.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.14.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.14.ffn_norm
F32
F32
[1280]
blk.14.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.14.ssm_a
F32
F32
[1, 24]
blk.14.ssm_conv1d.bias
F32
F32
[2048]
blk.14.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.14.ssm_d
F32
F32
[1, 24]
blk.14.ssm_dt.bias
F32
F32
[24]
blk.14.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.14.ssm_norm.weight
F32
F32
[1536]
blk.14.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.15
blk.15.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.15.attn_norm.weight
F32
F32
[1280]
blk.15.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.15.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.15.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.15.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.15.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.15.ffn_norm
F32
F32
[1280]
blk.15.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.15.ssm_a
F32
F32
[1, 24]
blk.15.ssm_conv1d.bias
F32
F32
[2048]
blk.15.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.15.ssm_d
F32
F32
[1, 24]
blk.15.ssm_dt.bias
F32
F32
[24]
blk.15.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.15.ssm_norm.weight
F32
F32
[1536]
blk.15.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.16
blk.16.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.16.attn_norm.weight
F32
F32
[1280]
blk.16.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.16.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.16.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.16.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.16.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.16.ffn_norm
F32
F32
[1280]
blk.16.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.16.ssm_a
F32
F32
[1, 24]
blk.16.ssm_conv1d.bias
F32
F32
[2048]
blk.16.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.16.ssm_d
F32
F32
[1, 24]
blk.16.ssm_dt.bias
F32
F32
[24]
blk.16.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.16.ssm_norm.weight
F32
F32
[1536]
blk.16.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.17
blk.17.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.17.attn_norm.weight
F32
F32
[1280]
blk.17.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.17.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.17.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.17.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.17.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.17.ffn_norm
F32
F32
[1280]
blk.17.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.17.ssm_a
F32
F32
[1, 24]
blk.17.ssm_conv1d.bias
F32
F32
[2048]
blk.17.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.17.ssm_d
F32
F32
[1, 24]
blk.17.ssm_dt.bias
F32
F32
[24]
blk.17.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.17.ssm_norm.weight
F32
F32
[1536]
blk.17.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.18
blk.18.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.18.attn_norm.weight
F32
F32
[1280]
blk.18.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.18.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.18.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.18.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.18.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.18.ffn_norm
F32
F32
[1280]
blk.18.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.18.ssm_a
F32
F32
[1, 24]
blk.18.ssm_conv1d.bias
F32
F32
[2048]
blk.18.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.18.ssm_d
F32
F32
[1, 24]
blk.18.ssm_dt.bias
F32
F32
[24]
blk.18.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.18.ssm_norm.weight
F32
F32
[1536]
blk.18.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.19
blk.19.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.19.attn_norm.weight
F32
F32
[1280]
blk.19.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.19.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.19.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.19.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.19.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.19.ffn_norm
F32
F32
[1280]
blk.19.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.19.ssm_a
F32
F32
[1, 24]
blk.19.ssm_conv1d.bias
F32
F32
[2048]
blk.19.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.19.ssm_d
F32
F32
[1, 24]
blk.19.ssm_dt.bias
F32
F32
[24]
blk.19.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.19.ssm_norm.weight
F32
F32
[1536]
blk.19.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.20
blk.20.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.20.attn_norm.weight
F32
F32
[1280]
blk.20.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.20.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.20.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.20.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.20.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.20.ffn_norm
F32
F32
[1280]
blk.20.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.20.ssm_a
F32
F32
[1, 24]
blk.20.ssm_conv1d.bias
F32
F32
[2048]
blk.20.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.20.ssm_d
F32
F32
[1, 24]
blk.20.ssm_dt.bias
F32
F32
[24]
blk.20.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.20.ssm_norm.weight
F32
F32
[1536]
blk.20.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.21
blk.21.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.21.attn_norm.weight
F32
F32
[1280]
blk.21.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.21.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.21.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.21.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.21.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.21.ffn_norm
F32
F32
[1280]
blk.21.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.21.ssm_a
F32
F32
[1, 24]
blk.21.ssm_conv1d.bias
F32
F32
[2048]
blk.21.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.21.ssm_d
F32
F32
[1, 24]
blk.21.ssm_dt.bias
F32
F32
[24]
blk.21.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.21.ssm_norm.weight
F32
F32
[1536]
blk.21.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.22
blk.22.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.22.attn_norm.weight
F32
F32
[1280]
blk.22.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.22.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.22.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.22.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.22.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.22.ffn_norm
F32
F32
[1280]
blk.22.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.22.ssm_a
F32
F32
[1, 24]
blk.22.ssm_conv1d.bias
F32
F32
[2048]
blk.22.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.22.ssm_d
F32
F32
[1, 24]
blk.22.ssm_dt.bias
F32
F32
[24]
blk.22.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.22.ssm_norm.weight
F32
F32
[1536]
blk.22.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.23
blk.23.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.23.attn_norm.weight
F32
F32
[1280]
blk.23.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.23.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.23.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.23.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.23.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.23.ffn_norm
F32
F32
[1280]
blk.23.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.23.ssm_a
F32
F32
[1, 24]
blk.23.ssm_conv1d.bias
F32
F32
[2048]
blk.23.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.23.ssm_d
F32
F32
[1, 24]
blk.23.ssm_dt.bias
F32
F32
[24]
blk.23.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.23.ssm_norm.weight
F32
F32
[1536]
blk.23.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.24
blk.24.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.24.attn_norm.weight
F32
F32
[1280]
blk.24.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.24.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.24.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.24.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.24.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.24.ffn_norm
F32
F32
[1280]
blk.24.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.24.ssm_a
F32
F32
[1, 24]
blk.24.ssm_conv1d.bias
F32
F32
[2048]
blk.24.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.24.ssm_d
F32
F32
[1, 24]
blk.24.ssm_dt.bias
F32
F32
[24]
blk.24.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.24.ssm_norm.weight
F32
F32
[1536]
blk.24.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.25
blk.25.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.25.attn_norm.weight
F32
F32
[1280]
blk.25.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.25.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.25.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.25.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.25.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.25.ffn_norm
F32
F32
[1280]
blk.25.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.25.ssm_a
F32
F32
[1, 24]
blk.25.ssm_conv1d.bias
F32
F32
[2048]
blk.25.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.25.ssm_d
F32
F32
[1, 24]
blk.25.ssm_dt.bias
F32
F32
[24]
blk.25.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.25.ssm_norm.weight
F32
F32
[1536]
blk.25.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.26
blk.26.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.26.attn_norm.weight
F32
F32
[1280]
blk.26.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.26.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.26.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.26.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.26.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.26.ffn_norm
F32
F32
[1280]
blk.26.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.26.ssm_a
F32
F32
[1, 24]
blk.26.ssm_conv1d.bias
F32
F32
[2048]
blk.26.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.26.ssm_d
F32
F32
[1, 24]
blk.26.ssm_dt.bias
F32
F32
[24]
blk.26.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.26.ssm_norm.weight
F32
F32
[1536]
blk.26.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.27
blk.27.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.27.attn_norm.weight
F32
F32
[1280]
blk.27.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.27.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.27.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.27.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.27.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.27.ffn_norm
F32
F32
[1280]
blk.27.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.27.ssm_a
F32
F32
[1, 24]
blk.27.ssm_conv1d.bias
F32
F32
[2048]
blk.27.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.27.ssm_d
F32
F32
[1, 24]
blk.27.ssm_dt.bias
F32
F32
[24]
blk.27.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.27.ssm_norm.weight
F32
F32
[1536]
blk.27.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.28
blk.28.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.28.attn_norm.weight
F32
F32
[1280]
blk.28.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.28.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.28.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.28.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.28.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.28.ffn_norm
F32
F32
[1280]
blk.28.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.28.ssm_a
F32
F32
[1, 24]
blk.28.ssm_conv1d.bias
F32
F32
[2048]
blk.28.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.28.ssm_d
F32
F32
[1, 24]
blk.28.ssm_dt.bias
F32
F32
[24]
blk.28.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.28.ssm_norm.weight
F32
F32
[1536]
blk.28.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.29
blk.29.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.29.attn_norm.weight
F32
F32
[1280]
blk.29.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.29.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.29.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.29.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.29.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.29.ffn_norm
F32
F32
[1280]
blk.29.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.29.ssm_a
F32
F32
[1, 24]
blk.29.ssm_conv1d.bias
F32
F32
[2048]
blk.29.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.29.ssm_d
F32
F32
[1, 24]
blk.29.ssm_dt.bias
F32
F32
[24]
blk.29.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.29.ssm_norm.weight
F32
F32
[1536]
blk.29.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.30
blk.30.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.30.attn_norm.weight
F32
F32
[1280]
blk.30.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.30.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.30.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.30.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.30.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.30.ffn_norm
F32
F32
[1280]
blk.30.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.30.ssm_a
F32
F32
[1, 24]
blk.30.ssm_conv1d.bias
F32
F32
[2048]
blk.30.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.30.ssm_d
F32
F32
[1, 24]
blk.30.ssm_dt.bias
F32
F32
[24]
blk.30.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.30.ssm_norm.weight
F32
F32
[1536]
blk.30.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.31
blk.31.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.31.attn_norm.weight
F32
F32
[1280]
blk.31.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.31.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.31.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.31.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.31.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.31.ffn_norm
F32
F32
[1280]
blk.31.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.31.ssm_a
F32
F32
[1, 24]
blk.31.ssm_conv1d.bias
F32
F32
[2048]
blk.31.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.31.ssm_d
F32
F32
[1, 24]
blk.31.ssm_dt.bias
F32
F32
[24]
blk.31.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.31.ssm_norm.weight
F32
F32
[1536]
blk.31.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.32
blk.32.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.32.attn_norm.weight
F32
F32
[1280]
blk.32.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.32.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.32.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.32.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.32.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.32.ffn_norm
F32
F32
[1280]
blk.32.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.32.ssm_a
F32
F32
[1, 24]
blk.32.ssm_conv1d.bias
F32
F32
[2048]
blk.32.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.32.ssm_d
F32
F32
[1, 24]
blk.32.ssm_dt.bias
F32
F32
[24]
blk.32.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.32.ssm_norm.weight
F32
F32
[1536]
blk.32.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.33
blk.33.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.33.attn_norm.weight
F32
F32
[1280]
blk.33.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.33.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.33.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.33.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.33.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.33.ffn_norm
F32
F32
[1280]
blk.33.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.33.ssm_a
F32
F32
[1, 24]
blk.33.ssm_conv1d.bias
F32
F32
[2048]
blk.33.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.33.ssm_d
F32
F32
[1, 24]
blk.33.ssm_dt.bias
F32
F32
[24]
blk.33.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.33.ssm_norm.weight
F32
F32
[1536]
blk.33.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.34
blk.34.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.34.attn_norm.weight
F32
F32
[1280]
blk.34.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.34.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.34.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.34.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.34.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.34.ffn_norm
F32
F32
[1280]
blk.34.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.34.ssm_a
F32
F32
[1, 24]
blk.34.ssm_conv1d.bias
F32
F32
[2048]
blk.34.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.34.ssm_d
F32
F32
[1, 24]
blk.34.ssm_dt.bias
F32
F32
[24]
blk.34.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.34.ssm_norm.weight
F32
F32
[1536]
blk.34.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.35
blk.35.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.35.attn_norm.weight
F32
F32
[1280]
blk.35.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.35.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.35.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.35.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.35.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.35.ffn_norm
F32
F32
[1280]
blk.35.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.35.ssm_a
F32
F32
[1, 24]
blk.35.ssm_conv1d.bias
F32
F32
[2048]
blk.35.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.35.ssm_d
F32
F32
[1, 24]
blk.35.ssm_dt.bias
F32
F32
[24]
blk.35.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.35.ssm_norm.weight
F32
F32
[1536]
blk.35.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.36
blk.36.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.36.attn_norm.weight
F32
F32
[1280]
blk.36.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.36.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.36.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.36.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.36.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.36.ffn_norm
F32
F32
[1280]
blk.36.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.36.ssm_a
F32
F32
[1, 24]
blk.36.ssm_conv1d.bias
F32
F32
[2048]
blk.36.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.36.ssm_d
F32
F32
[1, 24]
blk.36.ssm_dt.bias
F32
F32
[24]
blk.36.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.36.ssm_norm.weight
F32
F32
[1536]
blk.36.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.37
blk.37.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.37.attn_norm.weight
F32
F32
[1280]
blk.37.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.37.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.37.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.37.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.37.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.37.ffn_norm
F32
F32
[1280]
blk.37.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.37.ssm_a
F32
F32
[1, 24]
blk.37.ssm_conv1d.bias
F32
F32
[2048]
blk.37.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.37.ssm_d
F32
F32
[1, 24]
blk.37.ssm_dt.bias
F32
F32
[24]
blk.37.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.37.ssm_norm.weight
F32
F32
[1536]
blk.37.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.38
blk.38.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.38.attn_norm.weight
F32
F32
[1280]
blk.38.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.38.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.38.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.38.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.38.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.38.ffn_norm
F32
F32
[1280]
blk.38.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.38.ssm_a
F32
F32
[1, 24]
blk.38.ssm_conv1d.bias
F32
F32
[2048]
blk.38.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.38.ssm_d
F32
F32
[1, 24]
blk.38.ssm_dt.bias
F32
F32
[24]
blk.38.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.38.ssm_norm.weight
F32
F32
[1536]
blk.38.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.39
blk.39.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.39.attn_norm.weight
F32
F32
[1280]
blk.39.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.39.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.39.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.39.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.39.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.39.ffn_norm
F32
F32
[1280]
blk.39.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.39.ssm_a
F32
F32
[1, 24]
blk.39.ssm_conv1d.bias
F32
F32
[2048]
blk.39.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.39.ssm_d
F32
F32
[1, 24]
blk.39.ssm_dt.bias
F32
F32
[24]
blk.39.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.39.ssm_norm.weight
F32
F32
[1536]
blk.39.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.40
blk.40.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.40.attn_norm.weight
F32
F32
[1280]
blk.40.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.40.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.40.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.40.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.40.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.40.ffn_norm
F32
F32
[1280]
blk.40.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.40.ssm_a
F32
F32
[1, 24]
blk.40.ssm_conv1d.bias
F32
F32
[2048]
blk.40.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.40.ssm_d
F32
F32
[1, 24]
blk.40.ssm_dt.bias
F32
F32
[24]
blk.40.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.40.ssm_norm.weight
F32
F32
[1536]
blk.40.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.41
blk.41.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.41.attn_norm.weight
F32
F32
[1280]
blk.41.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.41.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.41.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.41.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.41.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.41.ffn_norm
F32
F32
[1280]
blk.41.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.41.ssm_a
F32
F32
[1, 24]
blk.41.ssm_conv1d.bias
F32
F32
[2048]
blk.41.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.41.ssm_d
F32
F32
[1, 24]
blk.41.ssm_dt.bias
F32
F32
[24]
blk.41.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.41.ssm_norm.weight
F32
F32
[1536]
blk.41.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.42
blk.42.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.42.attn_norm.weight
F32
F32
[1280]
blk.42.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.42.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.42.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.42.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.42.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.42.ffn_norm
F32
F32
[1280]
blk.42.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.42.ssm_a
F32
F32
[1, 24]
blk.42.ssm_conv1d.bias
F32
F32
[2048]
blk.42.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.42.ssm_d
F32
F32
[1, 24]
blk.42.ssm_dt.bias
F32
F32
[24]
blk.42.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.42.ssm_norm.weight
F32
F32
[1536]
blk.42.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.43
blk.43.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.43.attn_norm.weight
F32
F32
[1280]
blk.43.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.43.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.43.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.43.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.43.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.43.ffn_norm
F32
F32
[1280]
blk.43.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.43.ssm_a
F32
F32
[1, 24]
blk.43.ssm_conv1d.bias
F32
F32
[2048]
blk.43.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.43.ssm_d
F32
F32
[1, 24]
blk.43.ssm_dt.bias
F32
F32
[24]
blk.43.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.43.ssm_norm.weight
F32
F32
[1536]
blk.43.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.44
blk.44.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.44.attn_norm.weight
F32
F32
[1280]
blk.44.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.44.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.44.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.44.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.44.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.44.ffn_norm
F32
F32
[1280]
blk.44.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.44.ssm_a
F32
F32
[1, 24]
blk.44.ssm_conv1d.bias
F32
F32
[2048]
blk.44.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.44.ssm_d
F32
F32
[1, 24]
blk.44.ssm_dt.bias
F32
F32
[24]
blk.44.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.44.ssm_norm.weight
F32
F32
[1536]
blk.44.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.45
blk.45.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.45.attn_norm.weight
F32
F32
[1280]
blk.45.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.45.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.45.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.45.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.45.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.45.ffn_norm
F32
F32
[1280]
blk.45.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.45.ssm_a
F32
F32
[1, 24]
blk.45.ssm_conv1d.bias
F32
F32
[2048]
blk.45.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.45.ssm_d
F32
F32
[1, 24]
blk.45.ssm_dt.bias
F32
F32
[24]
blk.45.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.45.ssm_norm.weight
F32
F32
[1536]
blk.45.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.46
blk.46.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.46.attn_norm.weight
F32
F32
[1280]
blk.46.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.46.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.46.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.46.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.46.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.46.ffn_norm
F32
F32
[1280]
blk.46.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.46.ssm_a
F32
F32
[1, 24]
blk.46.ssm_conv1d.bias
F32
F32
[2048]
blk.46.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.46.ssm_d
F32
F32
[1, 24]
blk.46.ssm_dt.bias
F32
F32
[24]
blk.46.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.46.ssm_norm.weight
F32
F32
[1536]
blk.46.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.47
blk.47.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.47.attn_norm.weight
F32
F32
[1280]
blk.47.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.47.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.47.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.47.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.47.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.47.ffn_norm
F32
F32
[1280]
blk.47.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.47.ssm_a
F32
F32
[1, 24]
blk.47.ssm_conv1d.bias
F32
F32
[2048]
blk.47.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.47.ssm_d
F32
F32
[1, 24]
blk.47.ssm_dt.bias
F32
F32
[24]
blk.47.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.47.ssm_norm.weight
F32
F32
[1536]
blk.47.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.48
blk.48.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.48.attn_norm.weight
F32
F32
[1280]
blk.48.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.48.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.48.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.48.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.48.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.48.ffn_norm
F32
F32
[1280]
blk.48.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.48.ssm_a
F32
F32
[1, 24]
blk.48.ssm_conv1d.bias
F32
F32
[2048]
blk.48.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.48.ssm_d
F32
F32
[1, 24]
blk.48.ssm_dt.bias
F32
F32
[24]
blk.48.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.48.ssm_norm.weight
F32
F32
[1536]
blk.48.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.49
blk.49.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.49.attn_norm.weight
F32
F32
[1280]
blk.49.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.49.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.49.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.49.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.49.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.49.ffn_norm
F32
F32
[1280]
blk.49.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.49.ssm_a
F32
F32
[1, 24]
blk.49.ssm_conv1d.bias
F32
F32
[2048]
blk.49.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.49.ssm_d
F32
F32
[1, 24]
blk.49.ssm_dt.bias
F32
F32
[24]
blk.49.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.49.ssm_norm.weight
F32
F32
[1536]
blk.49.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.50
blk.50.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.50.attn_norm.weight
F32
F32
[1280]
blk.50.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.50.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.50.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.50.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.50.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.50.ffn_norm
F32
F32
[1280]
blk.50.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.50.ssm_a
F32
F32
[1, 24]
blk.50.ssm_conv1d.bias
F32
F32
[2048]
blk.50.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.50.ssm_d
F32
F32
[1, 24]
blk.50.ssm_dt.bias
F32
F32
[24]
blk.50.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.50.ssm_norm.weight
F32
F32
[1536]
blk.50.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.51
blk.51.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.51.attn_norm.weight
F32
F32
[1280]
blk.51.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.51.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.51.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.51.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.51.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.51.ffn_norm
F32
F32
[1280]
blk.51.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.51.ssm_a
F32
F32
[1, 24]
blk.51.ssm_conv1d.bias
F32
F32
[2048]
blk.51.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.51.ssm_d
F32
F32
[1, 24]
blk.51.ssm_dt.bias
F32
F32
[24]
blk.51.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.51.ssm_norm.weight
F32
F32
[1536]
blk.51.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.52
blk.52.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.52.attn_norm.weight
F32
F32
[1280]
blk.52.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.52.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.52.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.52.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.52.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.52.ffn_norm
F32
F32
[1280]
blk.52.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.52.ssm_a
F32
F32
[1, 24]
blk.52.ssm_conv1d.bias
F32
F32
[2048]
blk.52.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.52.ssm_d
F32
F32
[1, 24]
blk.52.ssm_dt.bias
F32
F32
[24]
blk.52.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.52.ssm_norm.weight
F32
F32
[1536]
blk.52.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.53
blk.53.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.53.attn_norm.weight
F32
F32
[1280]
blk.53.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.53.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.53.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.53.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.53.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.53.ffn_norm
F32
F32
[1280]
blk.53.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.53.ssm_a
F32
F32
[1, 24]
blk.53.ssm_conv1d.bias
F32
F32
[2048]
blk.53.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.53.ssm_d
F32
F32
[1, 24]
blk.53.ssm_dt.bias
F32
F32
[24]
blk.53.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.53.ssm_norm.weight
F32
F32
[1536]
blk.53.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.54
blk.54.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.54.attn_norm.weight
F32
F32
[1280]
blk.54.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.54.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.54.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.54.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.54.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.54.ffn_norm
F32
F32
[1280]
blk.54.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.54.ssm_a
F32
F32
[1, 24]
blk.54.ssm_conv1d.bias
F32
F32
[2048]
blk.54.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.54.ssm_d
F32
F32
[1, 24]
blk.54.ssm_dt.bias
F32
F32
[24]
blk.54.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.54.ssm_norm.weight
F32
F32
[1536]
blk.54.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.55
blk.55.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.55.attn_norm.weight
F32
F32
[1280]
blk.55.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.55.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.55.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.55.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.55.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.55.ffn_norm
F32
F32
[1280]
blk.55.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.55.ssm_a
F32
F32
[1, 24]
blk.55.ssm_conv1d.bias
F32
F32
[2048]
blk.55.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.55.ssm_d
F32
F32
[1, 24]
blk.55.ssm_dt.bias
F32
F32
[24]
blk.55.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.55.ssm_norm.weight
F32
F32
[1536]
blk.55.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.56
blk.56.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.56.attn_norm.weight
F32
F32
[1280]
blk.56.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.56.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.56.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.56.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.56.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.56.ffn_norm
F32
F32
[1280]
blk.56.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.56.ssm_a
F32
F32
[1, 24]
blk.56.ssm_conv1d.bias
F32
F32
[2048]
blk.56.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.56.ssm_d
F32
F32
[1, 24]
blk.56.ssm_dt.bias
F32
F32
[24]
blk.56.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.56.ssm_norm.weight
F32
F32
[1536]
blk.56.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.57
blk.57.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.57.attn_norm.weight
F32
F32
[1280]
blk.57.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.57.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.57.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.57.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.57.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.57.ffn_norm
F32
F32
[1280]
blk.57.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.57.ssm_a
F32
F32
[1, 24]
blk.57.ssm_conv1d.bias
F32
F32
[2048]
blk.57.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.57.ssm_d
F32
F32
[1, 24]
blk.57.ssm_dt.bias
F32
F32
[24]
blk.57.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.57.ssm_norm.weight
F32
F32
[1536]
blk.57.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.58
blk.58.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.58.attn_norm.weight
F32
F32
[1280]
blk.58.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.58.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.58.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.58.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.58.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.58.ffn_norm
F32
F32
[1280]
blk.58.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.58.ssm_a
F32
F32
[1, 24]
blk.58.ssm_conv1d.bias
F32
F32
[2048]
blk.58.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.58.ssm_d
F32
F32
[1, 24]
blk.58.ssm_dt.bias
F32
F32
[24]
blk.58.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.58.ssm_norm.weight
F32
F32
[1536]
blk.58.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.59
blk.59.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.59.attn_norm.weight
F32
F32
[1280]
blk.59.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.59.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.59.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.59.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.59.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.59.ffn_norm
F32
F32
[1280]
blk.59.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.59.ssm_a
F32
F32
[1, 24]
blk.59.ssm_conv1d.bias
F32
F32
[2048]
blk.59.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.59.ssm_d
F32
F32
[1, 24]
blk.59.ssm_dt.bias
F32
F32
[24]
blk.59.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.59.ssm_norm.weight
F32
F32
[1536]
blk.59.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.60
blk.60.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.60.attn_norm.weight
F32
F32
[1280]
blk.60.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.60.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.60.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.60.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.60.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.60.ffn_norm
F32
F32
[1280]
blk.60.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.60.ssm_a
F32
F32
[1, 24]
blk.60.ssm_conv1d.bias
F32
F32
[2048]
blk.60.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.60.ssm_d
F32
F32
[1, 24]
blk.60.ssm_dt.bias
F32
F32
[24]
blk.60.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.60.ssm_norm.weight
F32
F32
[1536]
blk.60.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.61
blk.61.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.61.attn_norm.weight
F32
F32
[1280]
blk.61.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.61.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.61.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.61.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.61.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.61.ffn_norm
F32
F32
[1280]
blk.61.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.61.ssm_a
F32
F32
[1, 24]
blk.61.ssm_conv1d.bias
F32
F32
[2048]
blk.61.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.61.ssm_d
F32
F32
[1, 24]
blk.61.ssm_dt.bias
F32
F32
[24]
blk.61.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.61.ssm_norm.weight
F32
F32
[1536]
blk.61.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.62
blk.62.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.62.attn_norm.weight
F32
F32
[1280]
blk.62.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.62.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.62.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.62.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.62.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.62.ffn_norm
F32
F32
[1280]
blk.62.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.62.ssm_a
F32
F32
[1, 24]
blk.62.ssm_conv1d.bias
F32
F32
[2048]
blk.62.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.62.ssm_d
F32
F32
[1, 24]
blk.62.ssm_dt.bias
F32
F32
[24]
blk.62.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.62.ssm_norm.weight
F32
F32
[1536]
blk.62.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.63
blk.63.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.63.attn_norm.weight
F32
F32
[1280]
blk.63.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.63.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.63.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.63.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.63.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.63.ffn_norm
F32
F32
[1280]
blk.63.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.63.ssm_a
F32
F32
[1, 24]
blk.63.ssm_conv1d.bias
F32
F32
[2048]
blk.63.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.63.ssm_d
F32
F32
[1, 24]
blk.63.ssm_dt.bias
F32
F32
[24]
blk.63.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.63.ssm_norm.weight
F32
F32
[1536]
blk.63.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.64
blk.64.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.64.attn_norm.weight
F32
F32
[1280]
blk.64.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.64.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.64.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.64.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.64.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.64.ffn_norm
F32
F32
[1280]
blk.64.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.64.ssm_a
F32
F32
[1, 24]
blk.64.ssm_conv1d.bias
F32
F32
[2048]
blk.64.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.64.ssm_d
F32
F32
[1, 24]
blk.64.ssm_dt.bias
F32
F32
[24]
blk.64.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.64.ssm_norm.weight
F32
F32
[1536]
blk.64.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
blk.65
blk.65.attn_k.weight
Q4_0
Q4_0
[1280, 256]
blk.65.attn_norm.weight
F32
F32
[1280]
blk.65.attn_output.weight
Q4_0
Q4_0
[768, 1280]
blk.65.attn_q.weight
Q4_0
Q4_0
[1280, 768]
blk.65.attn_v.weight
Q4_0
Q4_0
[1280, 256]
blk.65.ffn_down.weight
Q4_0
Q4_0
[3072, 1280]
blk.65.ffn_gate.weight
Q4_0
Q4_0
[1280, 3072]
blk.65.ffn_norm
F32
F32
[1280]
blk.65.ffn_up.weight
Q4_0
Q4_0
[1280, 3072]
blk.65.ssm_a
F32
F32
[1, 24]
blk.65.ssm_conv1d.bias
F32
F32
[2048]
blk.65.ssm_conv1d.weight
F32
F32
[4, 2048]
blk.65.ssm_d
F32
F32
[1, 24]
blk.65.ssm_dt.bias
F32
F32
[24]
blk.65.ssm_in.weight
Q4_0
Q4_0
[1280, 3608]
blk.65.ssm_norm.weight
F32
F32
[1536]
blk.65.ssm_out.weight
Q4_0
Q4_0
[1536, 1280]
output.weight
Q6_K
Q6_K
[1280, 65537]
output_norm.weight
F32
F32
[1280]