Models
GitHub
Discord
Turbo
Sign in
Download
Models
Download
GitHub
Discord
Sign in
Hudson
/
mamba-chat
:latest
205
Downloads
Updated
10 months ago
The first chat language model based on a state-space model architecture!
The first chat language model based on a state-space model architecture!
Cancel
mamba-chat:latest
...
/
model
720c4865c9d1 · 5.6GB
Metadata
general.architecture
mamba
mamba
general.file_type
F16
F16
mamba.attention.head_count
0
0
mamba.attention.layer_norm_rms_epsilon
1e-05
1e-05
mamba.block_count
64
64
mamba.context_length
1048576
1048576
mamba.embedding_length
2560
2560
mamba.feed_forward_length
0
0
mamba.ssm.conv_kernel
4
4
mamba.ssm.dt_b_c_rms
false
false
mamba.ssm.inner_size
5120
5120
mamba.ssm.state_size
16
16
mamba.ssm.time_step_rank
160
160
tokenizer.ggml.bos_token_id
0
0
tokenizer.ggml.eos_token_id
0
0
tokenizer.ggml.merges
[Ġ Ġ, Ġ t, Ġ a, h e, i n, ...]
[Ġ Ġ, Ġ t, Ġ a, h e, i n, ...]
tokenizer.ggml.model
gpt2
gpt2
tokenizer.ggml.padding_token_id
0
0
tokenizer.ggml.pre
olmo
olmo
tokenizer.ggml.token_type
[3, 3, 1, 1, 1, ...]
[3, 3, 1, 1, 1, ...]
tokenizer.ggml.tokens
[<|endoftext|>, <|padding|>, !, ", #, ...]
[<|endoftext|>, <|padding|>, !, ", #, ...]
tokenizer.ggml.unknown_token_id
0
0
Tensor
Name
Type
Shape
token_embd.weight
F16
F16
[2560, 50280]
blk.0
blk.0.attn_norm.weight
F32
F32
[2560]
blk.0.ssm_a
F32
F32
[16, 5120]
blk.0.ssm_conv1d.bias
F32
F32
[5120]
blk.0.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.0.ssm_d
F32
F32
[5120]
blk.0.ssm_dt.bias
F32
F32
[5120]
blk.0.ssm_dt.weight
F16
F16
[160, 5120]
blk.0.ssm_in.weight
F16
F16
[2560, 10240]
blk.0.ssm_out.weight
F16
F16
[5120, 2560]
blk.0.ssm_x.weight
F16
F16
[5120, 192]
blk.1
blk.1.attn_norm.weight
F32
F32
[2560]
blk.1.ssm_a
F32
F32
[16, 5120]
blk.1.ssm_conv1d.bias
F32
F32
[5120]
blk.1.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.1.ssm_d
F32
F32
[5120]
blk.1.ssm_dt.bias
F32
F32
[5120]
blk.1.ssm_dt.weight
F16
F16
[160, 5120]
blk.1.ssm_in.weight
F16
F16
[2560, 10240]
blk.1.ssm_out.weight
F16
F16
[5120, 2560]
blk.1.ssm_x.weight
F16
F16
[5120, 192]
blk.2
blk.2.attn_norm.weight
F32
F32
[2560]
blk.2.ssm_a
F32
F32
[16, 5120]
blk.2.ssm_conv1d.bias
F32
F32
[5120]
blk.2.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.2.ssm_d
F32
F32
[5120]
blk.2.ssm_dt.bias
F32
F32
[5120]
blk.2.ssm_dt.weight
F16
F16
[160, 5120]
blk.2.ssm_in.weight
F16
F16
[2560, 10240]
blk.2.ssm_out.weight
F16
F16
[5120, 2560]
blk.2.ssm_x.weight
F16
F16
[5120, 192]
blk.3
blk.3.attn_norm.weight
F32
F32
[2560]
blk.3.ssm_a
F32
F32
[16, 5120]
blk.3.ssm_conv1d.bias
F32
F32
[5120]
blk.3.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.3.ssm_d
F32
F32
[5120]
blk.3.ssm_dt.bias
F32
F32
[5120]
blk.3.ssm_dt.weight
F16
F16
[160, 5120]
blk.3.ssm_in.weight
F16
F16
[2560, 10240]
blk.3.ssm_out.weight
F16
F16
[5120, 2560]
blk.3.ssm_x.weight
F16
F16
[5120, 192]
blk.4
blk.4.attn_norm.weight
F32
F32
[2560]
blk.4.ssm_a
F32
F32
[16, 5120]
blk.4.ssm_conv1d.bias
F32
F32
[5120]
blk.4.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.4.ssm_d
F32
F32
[5120]
blk.4.ssm_dt.bias
F32
F32
[5120]
blk.4.ssm_dt.weight
F16
F16
[160, 5120]
blk.4.ssm_in.weight
F16
F16
[2560, 10240]
blk.4.ssm_out.weight
F16
F16
[5120, 2560]
blk.4.ssm_x.weight
F16
F16
[5120, 192]
blk.5
blk.5.attn_norm.weight
F32
F32
[2560]
blk.5.ssm_a
F32
F32
[16, 5120]
blk.5.ssm_conv1d.bias
F32
F32
[5120]
blk.5.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.5.ssm_d
F32
F32
[5120]
blk.5.ssm_dt.bias
F32
F32
[5120]
blk.5.ssm_dt.weight
F16
F16
[160, 5120]
blk.5.ssm_in.weight
F16
F16
[2560, 10240]
blk.5.ssm_out.weight
F16
F16
[5120, 2560]
blk.5.ssm_x.weight
F16
F16
[5120, 192]
blk.6
blk.6.attn_norm.weight
F32
F32
[2560]
blk.6.ssm_a
F32
F32
[16, 5120]
blk.6.ssm_conv1d.bias
F32
F32
[5120]
blk.6.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.6.ssm_d
F32
F32
[5120]
blk.6.ssm_dt.bias
F32
F32
[5120]
blk.6.ssm_dt.weight
F16
F16
[160, 5120]
blk.6.ssm_in.weight
F16
F16
[2560, 10240]
blk.6.ssm_out.weight
F16
F16
[5120, 2560]
blk.6.ssm_x.weight
F16
F16
[5120, 192]
blk.7
blk.7.attn_norm.weight
F32
F32
[2560]
blk.7.ssm_a
F32
F32
[16, 5120]
blk.7.ssm_conv1d.bias
F32
F32
[5120]
blk.7.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.7.ssm_d
F32
F32
[5120]
blk.7.ssm_dt.bias
F32
F32
[5120]
blk.7.ssm_dt.weight
F16
F16
[160, 5120]
blk.7.ssm_in.weight
F16
F16
[2560, 10240]
blk.7.ssm_out.weight
F16
F16
[5120, 2560]
blk.7.ssm_x.weight
F16
F16
[5120, 192]
blk.8
blk.8.attn_norm.weight
F32
F32
[2560]
blk.8.ssm_a
F32
F32
[16, 5120]
blk.8.ssm_conv1d.bias
F32
F32
[5120]
blk.8.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.8.ssm_d
F32
F32
[5120]
blk.8.ssm_dt.bias
F32
F32
[5120]
blk.8.ssm_dt.weight
F16
F16
[160, 5120]
blk.8.ssm_in.weight
F16
F16
[2560, 10240]
blk.8.ssm_out.weight
F16
F16
[5120, 2560]
blk.8.ssm_x.weight
F16
F16
[5120, 192]
blk.9
blk.9.attn_norm.weight
F32
F32
[2560]
blk.9.ssm_a
F32
F32
[16, 5120]
blk.9.ssm_conv1d.bias
F32
F32
[5120]
blk.9.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.9.ssm_d
F32
F32
[5120]
blk.9.ssm_dt.bias
F32
F32
[5120]
blk.9.ssm_dt.weight
F16
F16
[160, 5120]
blk.9.ssm_in.weight
F16
F16
[2560, 10240]
blk.9.ssm_out.weight
F16
F16
[5120, 2560]
blk.9.ssm_x.weight
F16
F16
[5120, 192]
blk.10
blk.10.attn_norm.weight
F32
F32
[2560]
blk.10.ssm_a
F32
F32
[16, 5120]
blk.10.ssm_conv1d.bias
F32
F32
[5120]
blk.10.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.10.ssm_d
F32
F32
[5120]
blk.10.ssm_dt.bias
F32
F32
[5120]
blk.10.ssm_dt.weight
F16
F16
[160, 5120]
blk.10.ssm_in.weight
F16
F16
[2560, 10240]
blk.10.ssm_out.weight
F16
F16
[5120, 2560]
blk.10.ssm_x.weight
F16
F16
[5120, 192]
blk.11
blk.11.attn_norm.weight
F32
F32
[2560]
blk.11.ssm_a
F32
F32
[16, 5120]
blk.11.ssm_conv1d.bias
F32
F32
[5120]
blk.11.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.11.ssm_d
F32
F32
[5120]
blk.11.ssm_dt.bias
F32
F32
[5120]
blk.11.ssm_dt.weight
F16
F16
[160, 5120]
blk.11.ssm_in.weight
F16
F16
[2560, 10240]
blk.11.ssm_out.weight
F16
F16
[5120, 2560]
blk.11.ssm_x.weight
F16
F16
[5120, 192]
blk.12
blk.12.attn_norm.weight
F32
F32
[2560]
blk.12.ssm_a
F32
F32
[16, 5120]
blk.12.ssm_conv1d.bias
F32
F32
[5120]
blk.12.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.12.ssm_d
F32
F32
[5120]
blk.12.ssm_dt.bias
F32
F32
[5120]
blk.12.ssm_dt.weight
F16
F16
[160, 5120]
blk.12.ssm_in.weight
F16
F16
[2560, 10240]
blk.12.ssm_out.weight
F16
F16
[5120, 2560]
blk.12.ssm_x.weight
F16
F16
[5120, 192]
blk.13
blk.13.attn_norm.weight
F32
F32
[2560]
blk.13.ssm_a
F32
F32
[16, 5120]
blk.13.ssm_conv1d.bias
F32
F32
[5120]
blk.13.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.13.ssm_d
F32
F32
[5120]
blk.13.ssm_dt.bias
F32
F32
[5120]
blk.13.ssm_dt.weight
F16
F16
[160, 5120]
blk.13.ssm_in.weight
F16
F16
[2560, 10240]
blk.13.ssm_out.weight
F16
F16
[5120, 2560]
blk.13.ssm_x.weight
F16
F16
[5120, 192]
blk.14
blk.14.attn_norm.weight
F32
F32
[2560]
blk.14.ssm_a
F32
F32
[16, 5120]
blk.14.ssm_conv1d.bias
F32
F32
[5120]
blk.14.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.14.ssm_d
F32
F32
[5120]
blk.14.ssm_dt.bias
F32
F32
[5120]
blk.14.ssm_dt.weight
F16
F16
[160, 5120]
blk.14.ssm_in.weight
F16
F16
[2560, 10240]
blk.14.ssm_out.weight
F16
F16
[5120, 2560]
blk.14.ssm_x.weight
F16
F16
[5120, 192]
blk.15
blk.15.attn_norm.weight
F32
F32
[2560]
blk.15.ssm_a
F32
F32
[16, 5120]
blk.15.ssm_conv1d.bias
F32
F32
[5120]
blk.15.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.15.ssm_d
F32
F32
[5120]
blk.15.ssm_dt.bias
F32
F32
[5120]
blk.15.ssm_dt.weight
F16
F16
[160, 5120]
blk.15.ssm_in.weight
F16
F16
[2560, 10240]
blk.15.ssm_out.weight
F16
F16
[5120, 2560]
blk.15.ssm_x.weight
F16
F16
[5120, 192]
blk.16
blk.16.attn_norm.weight
F32
F32
[2560]
blk.16.ssm_a
F32
F32
[16, 5120]
blk.16.ssm_conv1d.bias
F32
F32
[5120]
blk.16.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.16.ssm_d
F32
F32
[5120]
blk.16.ssm_dt.bias
F32
F32
[5120]
blk.16.ssm_dt.weight
F16
F16
[160, 5120]
blk.16.ssm_in.weight
F16
F16
[2560, 10240]
blk.16.ssm_out.weight
F16
F16
[5120, 2560]
blk.16.ssm_x.weight
F16
F16
[5120, 192]
blk.17
blk.17.attn_norm.weight
F32
F32
[2560]
blk.17.ssm_a
F32
F32
[16, 5120]
blk.17.ssm_conv1d.bias
F32
F32
[5120]
blk.17.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.17.ssm_d
F32
F32
[5120]
blk.17.ssm_dt.bias
F32
F32
[5120]
blk.17.ssm_dt.weight
F16
F16
[160, 5120]
blk.17.ssm_in.weight
F16
F16
[2560, 10240]
blk.17.ssm_out.weight
F16
F16
[5120, 2560]
blk.17.ssm_x.weight
F16
F16
[5120, 192]
blk.18
blk.18.attn_norm.weight
F32
F32
[2560]
blk.18.ssm_a
F32
F32
[16, 5120]
blk.18.ssm_conv1d.bias
F32
F32
[5120]
blk.18.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.18.ssm_d
F32
F32
[5120]
blk.18.ssm_dt.bias
F32
F32
[5120]
blk.18.ssm_dt.weight
F16
F16
[160, 5120]
blk.18.ssm_in.weight
F16
F16
[2560, 10240]
blk.18.ssm_out.weight
F16
F16
[5120, 2560]
blk.18.ssm_x.weight
F16
F16
[5120, 192]
blk.19
blk.19.attn_norm.weight
F32
F32
[2560]
blk.19.ssm_a
F32
F32
[16, 5120]
blk.19.ssm_conv1d.bias
F32
F32
[5120]
blk.19.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.19.ssm_d
F32
F32
[5120]
blk.19.ssm_dt.bias
F32
F32
[5120]
blk.19.ssm_dt.weight
F16
F16
[160, 5120]
blk.19.ssm_in.weight
F16
F16
[2560, 10240]
blk.19.ssm_out.weight
F16
F16
[5120, 2560]
blk.19.ssm_x.weight
F16
F16
[5120, 192]
blk.20
blk.20.attn_norm.weight
F32
F32
[2560]
blk.20.ssm_a
F32
F32
[16, 5120]
blk.20.ssm_conv1d.bias
F32
F32
[5120]
blk.20.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.20.ssm_d
F32
F32
[5120]
blk.20.ssm_dt.bias
F32
F32
[5120]
blk.20.ssm_dt.weight
F16
F16
[160, 5120]
blk.20.ssm_in.weight
F16
F16
[2560, 10240]
blk.20.ssm_out.weight
F16
F16
[5120, 2560]
blk.20.ssm_x.weight
F16
F16
[5120, 192]
blk.21
blk.21.attn_norm.weight
F32
F32
[2560]
blk.21.ssm_a
F32
F32
[16, 5120]
blk.21.ssm_conv1d.bias
F32
F32
[5120]
blk.21.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.21.ssm_d
F32
F32
[5120]
blk.21.ssm_dt.bias
F32
F32
[5120]
blk.21.ssm_dt.weight
F16
F16
[160, 5120]
blk.21.ssm_in.weight
F16
F16
[2560, 10240]
blk.21.ssm_out.weight
F16
F16
[5120, 2560]
blk.21.ssm_x.weight
F16
F16
[5120, 192]
blk.22
blk.22.attn_norm.weight
F32
F32
[2560]
blk.22.ssm_a
F32
F32
[16, 5120]
blk.22.ssm_conv1d.bias
F32
F32
[5120]
blk.22.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.22.ssm_d
F32
F32
[5120]
blk.22.ssm_dt.bias
F32
F32
[5120]
blk.22.ssm_dt.weight
F16
F16
[160, 5120]
blk.22.ssm_in.weight
F16
F16
[2560, 10240]
blk.22.ssm_out.weight
F16
F16
[5120, 2560]
blk.22.ssm_x.weight
F16
F16
[5120, 192]
blk.23
blk.23.attn_norm.weight
F32
F32
[2560]
blk.23.ssm_a
F32
F32
[16, 5120]
blk.23.ssm_conv1d.bias
F32
F32
[5120]
blk.23.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.23.ssm_d
F32
F32
[5120]
blk.23.ssm_dt.bias
F32
F32
[5120]
blk.23.ssm_dt.weight
F16
F16
[160, 5120]
blk.23.ssm_in.weight
F16
F16
[2560, 10240]
blk.23.ssm_out.weight
F16
F16
[5120, 2560]
blk.23.ssm_x.weight
F16
F16
[5120, 192]
blk.24
blk.24.attn_norm.weight
F32
F32
[2560]
blk.24.ssm_a
F32
F32
[16, 5120]
blk.24.ssm_conv1d.bias
F32
F32
[5120]
blk.24.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.24.ssm_d
F32
F32
[5120]
blk.24.ssm_dt.bias
F32
F32
[5120]
blk.24.ssm_dt.weight
F16
F16
[160, 5120]
blk.24.ssm_in.weight
F16
F16
[2560, 10240]
blk.24.ssm_out.weight
F16
F16
[5120, 2560]
blk.24.ssm_x.weight
F16
F16
[5120, 192]
blk.25
blk.25.attn_norm.weight
F32
F32
[2560]
blk.25.ssm_a
F32
F32
[16, 5120]
blk.25.ssm_conv1d.bias
F32
F32
[5120]
blk.25.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.25.ssm_d
F32
F32
[5120]
blk.25.ssm_dt.bias
F32
F32
[5120]
blk.25.ssm_dt.weight
F16
F16
[160, 5120]
blk.25.ssm_in.weight
F16
F16
[2560, 10240]
blk.25.ssm_out.weight
F16
F16
[5120, 2560]
blk.25.ssm_x.weight
F16
F16
[5120, 192]
blk.26
blk.26.attn_norm.weight
F32
F32
[2560]
blk.26.ssm_a
F32
F32
[16, 5120]
blk.26.ssm_conv1d.bias
F32
F32
[5120]
blk.26.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.26.ssm_d
F32
F32
[5120]
blk.26.ssm_dt.bias
F32
F32
[5120]
blk.26.ssm_dt.weight
F16
F16
[160, 5120]
blk.26.ssm_in.weight
F16
F16
[2560, 10240]
blk.26.ssm_out.weight
F16
F16
[5120, 2560]
blk.26.ssm_x.weight
F16
F16
[5120, 192]
blk.27
blk.27.attn_norm.weight
F32
F32
[2560]
blk.27.ssm_a
F32
F32
[16, 5120]
blk.27.ssm_conv1d.bias
F32
F32
[5120]
blk.27.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.27.ssm_d
F32
F32
[5120]
blk.27.ssm_dt.bias
F32
F32
[5120]
blk.27.ssm_dt.weight
F16
F16
[160, 5120]
blk.27.ssm_in.weight
F16
F16
[2560, 10240]
blk.27.ssm_out.weight
F16
F16
[5120, 2560]
blk.27.ssm_x.weight
F16
F16
[5120, 192]
blk.28
blk.28.attn_norm.weight
F32
F32
[2560]
blk.28.ssm_a
F32
F32
[16, 5120]
blk.28.ssm_conv1d.bias
F32
F32
[5120]
blk.28.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.28.ssm_d
F32
F32
[5120]
blk.28.ssm_dt.bias
F32
F32
[5120]
blk.28.ssm_dt.weight
F16
F16
[160, 5120]
blk.28.ssm_in.weight
F16
F16
[2560, 10240]
blk.28.ssm_out.weight
F16
F16
[5120, 2560]
blk.28.ssm_x.weight
F16
F16
[5120, 192]
blk.29
blk.29.attn_norm.weight
F32
F32
[2560]
blk.29.ssm_a
F32
F32
[16, 5120]
blk.29.ssm_conv1d.bias
F32
F32
[5120]
blk.29.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.29.ssm_d
F32
F32
[5120]
blk.29.ssm_dt.bias
F32
F32
[5120]
blk.29.ssm_dt.weight
F16
F16
[160, 5120]
blk.29.ssm_in.weight
F16
F16
[2560, 10240]
blk.29.ssm_out.weight
F16
F16
[5120, 2560]
blk.29.ssm_x.weight
F16
F16
[5120, 192]
blk.30
blk.30.attn_norm.weight
F32
F32
[2560]
blk.30.ssm_a
F32
F32
[16, 5120]
blk.30.ssm_conv1d.bias
F32
F32
[5120]
blk.30.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.30.ssm_d
F32
F32
[5120]
blk.30.ssm_dt.bias
F32
F32
[5120]
blk.30.ssm_dt.weight
F16
F16
[160, 5120]
blk.30.ssm_in.weight
F16
F16
[2560, 10240]
blk.30.ssm_out.weight
F16
F16
[5120, 2560]
blk.30.ssm_x.weight
F16
F16
[5120, 192]
blk.31
blk.31.attn_norm.weight
F32
F32
[2560]
blk.31.ssm_a
F32
F32
[16, 5120]
blk.31.ssm_conv1d.bias
F32
F32
[5120]
blk.31.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.31.ssm_d
F32
F32
[5120]
blk.31.ssm_dt.bias
F32
F32
[5120]
blk.31.ssm_dt.weight
F16
F16
[160, 5120]
blk.31.ssm_in.weight
F16
F16
[2560, 10240]
blk.31.ssm_out.weight
F16
F16
[5120, 2560]
blk.31.ssm_x.weight
F16
F16
[5120, 192]
blk.32
blk.32.attn_norm.weight
F32
F32
[2560]
blk.32.ssm_a
F32
F32
[16, 5120]
blk.32.ssm_conv1d.bias
F32
F32
[5120]
blk.32.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.32.ssm_d
F32
F32
[5120]
blk.32.ssm_dt.bias
F32
F32
[5120]
blk.32.ssm_dt.weight
F16
F16
[160, 5120]
blk.32.ssm_in.weight
F16
F16
[2560, 10240]
blk.32.ssm_out.weight
F16
F16
[5120, 2560]
blk.32.ssm_x.weight
F16
F16
[5120, 192]
blk.33
blk.33.attn_norm.weight
F32
F32
[2560]
blk.33.ssm_a
F32
F32
[16, 5120]
blk.33.ssm_conv1d.bias
F32
F32
[5120]
blk.33.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.33.ssm_d
F32
F32
[5120]
blk.33.ssm_dt.bias
F32
F32
[5120]
blk.33.ssm_dt.weight
F16
F16
[160, 5120]
blk.33.ssm_in.weight
F16
F16
[2560, 10240]
blk.33.ssm_out.weight
F16
F16
[5120, 2560]
blk.33.ssm_x.weight
F16
F16
[5120, 192]
blk.34
blk.34.attn_norm.weight
F32
F32
[2560]
blk.34.ssm_a
F32
F32
[16, 5120]
blk.34.ssm_conv1d.bias
F32
F32
[5120]
blk.34.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.34.ssm_d
F32
F32
[5120]
blk.34.ssm_dt.bias
F32
F32
[5120]
blk.34.ssm_dt.weight
F16
F16
[160, 5120]
blk.34.ssm_in.weight
F16
F16
[2560, 10240]
blk.34.ssm_out.weight
F16
F16
[5120, 2560]
blk.34.ssm_x.weight
F16
F16
[5120, 192]
blk.35
blk.35.attn_norm.weight
F32
F32
[2560]
blk.35.ssm_a
F32
F32
[16, 5120]
blk.35.ssm_conv1d.bias
F32
F32
[5120]
blk.35.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.35.ssm_d
F32
F32
[5120]
blk.35.ssm_dt.bias
F32
F32
[5120]
blk.35.ssm_dt.weight
F16
F16
[160, 5120]
blk.35.ssm_in.weight
F16
F16
[2560, 10240]
blk.35.ssm_out.weight
F16
F16
[5120, 2560]
blk.35.ssm_x.weight
F16
F16
[5120, 192]
blk.36
blk.36.attn_norm.weight
F32
F32
[2560]
blk.36.ssm_a
F32
F32
[16, 5120]
blk.36.ssm_conv1d.bias
F32
F32
[5120]
blk.36.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.36.ssm_d
F32
F32
[5120]
blk.36.ssm_dt.bias
F32
F32
[5120]
blk.36.ssm_dt.weight
F16
F16
[160, 5120]
blk.36.ssm_in.weight
F16
F16
[2560, 10240]
blk.36.ssm_out.weight
F16
F16
[5120, 2560]
blk.36.ssm_x.weight
F16
F16
[5120, 192]
blk.37
blk.37.attn_norm.weight
F32
F32
[2560]
blk.37.ssm_a
F32
F32
[16, 5120]
blk.37.ssm_conv1d.bias
F32
F32
[5120]
blk.37.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.37.ssm_d
F32
F32
[5120]
blk.37.ssm_dt.bias
F32
F32
[5120]
blk.37.ssm_dt.weight
F16
F16
[160, 5120]
blk.37.ssm_in.weight
F16
F16
[2560, 10240]
blk.37.ssm_out.weight
F16
F16
[5120, 2560]
blk.37.ssm_x.weight
F16
F16
[5120, 192]
blk.38
blk.38.attn_norm.weight
F32
F32
[2560]
blk.38.ssm_a
F32
F32
[16, 5120]
blk.38.ssm_conv1d.bias
F32
F32
[5120]
blk.38.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.38.ssm_d
F32
F32
[5120]
blk.38.ssm_dt.bias
F32
F32
[5120]
blk.38.ssm_dt.weight
F16
F16
[160, 5120]
blk.38.ssm_in.weight
F16
F16
[2560, 10240]
blk.38.ssm_out.weight
F16
F16
[5120, 2560]
blk.38.ssm_x.weight
F16
F16
[5120, 192]
blk.39
blk.39.attn_norm.weight
F32
F32
[2560]
blk.39.ssm_a
F32
F32
[16, 5120]
blk.39.ssm_conv1d.bias
F32
F32
[5120]
blk.39.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.39.ssm_d
F32
F32
[5120]
blk.39.ssm_dt.bias
F32
F32
[5120]
blk.39.ssm_dt.weight
F16
F16
[160, 5120]
blk.39.ssm_in.weight
F16
F16
[2560, 10240]
blk.39.ssm_out.weight
F16
F16
[5120, 2560]
blk.39.ssm_x.weight
F16
F16
[5120, 192]
blk.40
blk.40.attn_norm.weight
F32
F32
[2560]
blk.40.ssm_a
F32
F32
[16, 5120]
blk.40.ssm_conv1d.bias
F32
F32
[5120]
blk.40.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.40.ssm_d
F32
F32
[5120]
blk.40.ssm_dt.bias
F32
F32
[5120]
blk.40.ssm_dt.weight
F16
F16
[160, 5120]
blk.40.ssm_in.weight
F16
F16
[2560, 10240]
blk.40.ssm_out.weight
F16
F16
[5120, 2560]
blk.40.ssm_x.weight
F16
F16
[5120, 192]
blk.41
blk.41.attn_norm.weight
F32
F32
[2560]
blk.41.ssm_a
F32
F32
[16, 5120]
blk.41.ssm_conv1d.bias
F32
F32
[5120]
blk.41.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.41.ssm_d
F32
F32
[5120]
blk.41.ssm_dt.bias
F32
F32
[5120]
blk.41.ssm_dt.weight
F16
F16
[160, 5120]
blk.41.ssm_in.weight
F16
F16
[2560, 10240]
blk.41.ssm_out.weight
F16
F16
[5120, 2560]
blk.41.ssm_x.weight
F16
F16
[5120, 192]
blk.42
blk.42.attn_norm.weight
F32
F32
[2560]
blk.42.ssm_a
F32
F32
[16, 5120]
blk.42.ssm_conv1d.bias
F32
F32
[5120]
blk.42.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.42.ssm_d
F32
F32
[5120]
blk.42.ssm_dt.bias
F32
F32
[5120]
blk.42.ssm_dt.weight
F16
F16
[160, 5120]
blk.42.ssm_in.weight
F16
F16
[2560, 10240]
blk.42.ssm_out.weight
F16
F16
[5120, 2560]
blk.42.ssm_x.weight
F16
F16
[5120, 192]
blk.43
blk.43.attn_norm.weight
F32
F32
[2560]
blk.43.ssm_a
F32
F32
[16, 5120]
blk.43.ssm_conv1d.bias
F32
F32
[5120]
blk.43.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.43.ssm_d
F32
F32
[5120]
blk.43.ssm_dt.bias
F32
F32
[5120]
blk.43.ssm_dt.weight
F16
F16
[160, 5120]
blk.43.ssm_in.weight
F16
F16
[2560, 10240]
blk.43.ssm_out.weight
F16
F16
[5120, 2560]
blk.43.ssm_x.weight
F16
F16
[5120, 192]
blk.44
blk.44.attn_norm.weight
F32
F32
[2560]
blk.44.ssm_a
F32
F32
[16, 5120]
blk.44.ssm_conv1d.bias
F32
F32
[5120]
blk.44.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.44.ssm_d
F32
F32
[5120]
blk.44.ssm_dt.bias
F32
F32
[5120]
blk.44.ssm_dt.weight
F16
F16
[160, 5120]
blk.44.ssm_in.weight
F16
F16
[2560, 10240]
blk.44.ssm_out.weight
F16
F16
[5120, 2560]
blk.44.ssm_x.weight
F16
F16
[5120, 192]
blk.45
blk.45.attn_norm.weight
F32
F32
[2560]
blk.45.ssm_a
F32
F32
[16, 5120]
blk.45.ssm_conv1d.bias
F32
F32
[5120]
blk.45.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.45.ssm_d
F32
F32
[5120]
blk.45.ssm_dt.bias
F32
F32
[5120]
blk.45.ssm_dt.weight
F16
F16
[160, 5120]
blk.45.ssm_in.weight
F16
F16
[2560, 10240]
blk.45.ssm_out.weight
F16
F16
[5120, 2560]
blk.45.ssm_x.weight
F16
F16
[5120, 192]
blk.46
blk.46.attn_norm.weight
F32
F32
[2560]
blk.46.ssm_a
F32
F32
[16, 5120]
blk.46.ssm_conv1d.bias
F32
F32
[5120]
blk.46.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.46.ssm_d
F32
F32
[5120]
blk.46.ssm_dt.bias
F32
F32
[5120]
blk.46.ssm_dt.weight
F16
F16
[160, 5120]
blk.46.ssm_in.weight
F16
F16
[2560, 10240]
blk.46.ssm_out.weight
F16
F16
[5120, 2560]
blk.46.ssm_x.weight
F16
F16
[5120, 192]
blk.47
blk.47.attn_norm.weight
F32
F32
[2560]
blk.47.ssm_a
F32
F32
[16, 5120]
blk.47.ssm_conv1d.bias
F32
F32
[5120]
blk.47.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.47.ssm_d
F32
F32
[5120]
blk.47.ssm_dt.bias
F32
F32
[5120]
blk.47.ssm_dt.weight
F16
F16
[160, 5120]
blk.47.ssm_in.weight
F16
F16
[2560, 10240]
blk.47.ssm_out.weight
F16
F16
[5120, 2560]
blk.47.ssm_x.weight
F16
F16
[5120, 192]
blk.48
blk.48.attn_norm.weight
F32
F32
[2560]
blk.48.ssm_a
F32
F32
[16, 5120]
blk.48.ssm_conv1d.bias
F32
F32
[5120]
blk.48.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.48.ssm_d
F32
F32
[5120]
blk.48.ssm_dt.bias
F32
F32
[5120]
blk.48.ssm_dt.weight
F16
F16
[160, 5120]
blk.48.ssm_in.weight
F16
F16
[2560, 10240]
blk.48.ssm_out.weight
F16
F16
[5120, 2560]
blk.48.ssm_x.weight
F16
F16
[5120, 192]
blk.49
blk.49.attn_norm.weight
F32
F32
[2560]
blk.49.ssm_a
F32
F32
[16, 5120]
blk.49.ssm_conv1d.bias
F32
F32
[5120]
blk.49.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.49.ssm_d
F32
F32
[5120]
blk.49.ssm_dt.bias
F32
F32
[5120]
blk.49.ssm_dt.weight
F16
F16
[160, 5120]
blk.49.ssm_in.weight
F16
F16
[2560, 10240]
blk.49.ssm_out.weight
F16
F16
[5120, 2560]
blk.49.ssm_x.weight
F16
F16
[5120, 192]
blk.50
blk.50.attn_norm.weight
F32
F32
[2560]
blk.50.ssm_a
F32
F32
[16, 5120]
blk.50.ssm_conv1d.bias
F32
F32
[5120]
blk.50.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.50.ssm_d
F32
F32
[5120]
blk.50.ssm_dt.bias
F32
F32
[5120]
blk.50.ssm_dt.weight
F16
F16
[160, 5120]
blk.50.ssm_in.weight
F16
F16
[2560, 10240]
blk.50.ssm_out.weight
F16
F16
[5120, 2560]
blk.50.ssm_x.weight
F16
F16
[5120, 192]
blk.51
blk.51.attn_norm.weight
F32
F32
[2560]
blk.51.ssm_a
F32
F32
[16, 5120]
blk.51.ssm_conv1d.bias
F32
F32
[5120]
blk.51.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.51.ssm_d
F32
F32
[5120]
blk.51.ssm_dt.bias
F32
F32
[5120]
blk.51.ssm_dt.weight
F16
F16
[160, 5120]
blk.51.ssm_in.weight
F16
F16
[2560, 10240]
blk.51.ssm_out.weight
F16
F16
[5120, 2560]
blk.51.ssm_x.weight
F16
F16
[5120, 192]
blk.52
blk.52.attn_norm.weight
F32
F32
[2560]
blk.52.ssm_a
F32
F32
[16, 5120]
blk.52.ssm_conv1d.bias
F32
F32
[5120]
blk.52.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.52.ssm_d
F32
F32
[5120]
blk.52.ssm_dt.bias
F32
F32
[5120]
blk.52.ssm_dt.weight
F16
F16
[160, 5120]
blk.52.ssm_in.weight
F16
F16
[2560, 10240]
blk.52.ssm_out.weight
F16
F16
[5120, 2560]
blk.52.ssm_x.weight
F16
F16
[5120, 192]
blk.53
blk.53.attn_norm.weight
F32
F32
[2560]
blk.53.ssm_a
F32
F32
[16, 5120]
blk.53.ssm_conv1d.bias
F32
F32
[5120]
blk.53.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.53.ssm_d
F32
F32
[5120]
blk.53.ssm_dt.bias
F32
F32
[5120]
blk.53.ssm_dt.weight
F16
F16
[160, 5120]
blk.53.ssm_in.weight
F16
F16
[2560, 10240]
blk.53.ssm_out.weight
F16
F16
[5120, 2560]
blk.53.ssm_x.weight
F16
F16
[5120, 192]
blk.54
blk.54.attn_norm.weight
F32
F32
[2560]
blk.54.ssm_a
F32
F32
[16, 5120]
blk.54.ssm_conv1d.bias
F32
F32
[5120]
blk.54.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.54.ssm_d
F32
F32
[5120]
blk.54.ssm_dt.bias
F32
F32
[5120]
blk.54.ssm_dt.weight
F16
F16
[160, 5120]
blk.54.ssm_in.weight
F16
F16
[2560, 10240]
blk.54.ssm_out.weight
F16
F16
[5120, 2560]
blk.54.ssm_x.weight
F16
F16
[5120, 192]
blk.55
blk.55.attn_norm.weight
F32
F32
[2560]
blk.55.ssm_a
F32
F32
[16, 5120]
blk.55.ssm_conv1d.bias
F32
F32
[5120]
blk.55.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.55.ssm_d
F32
F32
[5120]
blk.55.ssm_dt.bias
F32
F32
[5120]
blk.55.ssm_dt.weight
F16
F16
[160, 5120]
blk.55.ssm_in.weight
F16
F16
[2560, 10240]
blk.55.ssm_out.weight
F16
F16
[5120, 2560]
blk.55.ssm_x.weight
F16
F16
[5120, 192]
blk.56
blk.56.attn_norm.weight
F32
F32
[2560]
blk.56.ssm_a
F32
F32
[16, 5120]
blk.56.ssm_conv1d.bias
F32
F32
[5120]
blk.56.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.56.ssm_d
F32
F32
[5120]
blk.56.ssm_dt.bias
F32
F32
[5120]
blk.56.ssm_dt.weight
F16
F16
[160, 5120]
blk.56.ssm_in.weight
F16
F16
[2560, 10240]
blk.56.ssm_out.weight
F16
F16
[5120, 2560]
blk.56.ssm_x.weight
F16
F16
[5120, 192]
blk.57
blk.57.attn_norm.weight
F32
F32
[2560]
blk.57.ssm_a
F32
F32
[16, 5120]
blk.57.ssm_conv1d.bias
F32
F32
[5120]
blk.57.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.57.ssm_d
F32
F32
[5120]
blk.57.ssm_dt.bias
F32
F32
[5120]
blk.57.ssm_dt.weight
F16
F16
[160, 5120]
blk.57.ssm_in.weight
F16
F16
[2560, 10240]
blk.57.ssm_out.weight
F16
F16
[5120, 2560]
blk.57.ssm_x.weight
F16
F16
[5120, 192]
blk.58
blk.58.attn_norm.weight
F32
F32
[2560]
blk.58.ssm_a
F32
F32
[16, 5120]
blk.58.ssm_conv1d.bias
F32
F32
[5120]
blk.58.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.58.ssm_d
F32
F32
[5120]
blk.58.ssm_dt.bias
F32
F32
[5120]
blk.58.ssm_dt.weight
F16
F16
[160, 5120]
blk.58.ssm_in.weight
F16
F16
[2560, 10240]
blk.58.ssm_out.weight
F16
F16
[5120, 2560]
blk.58.ssm_x.weight
F16
F16
[5120, 192]
blk.59
blk.59.attn_norm.weight
F32
F32
[2560]
blk.59.ssm_a
F32
F32
[16, 5120]
blk.59.ssm_conv1d.bias
F32
F32
[5120]
blk.59.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.59.ssm_d
F32
F32
[5120]
blk.59.ssm_dt.bias
F32
F32
[5120]
blk.59.ssm_dt.weight
F16
F16
[160, 5120]
blk.59.ssm_in.weight
F16
F16
[2560, 10240]
blk.59.ssm_out.weight
F16
F16
[5120, 2560]
blk.59.ssm_x.weight
F16
F16
[5120, 192]
blk.60
blk.60.attn_norm.weight
F32
F32
[2560]
blk.60.ssm_a
F32
F32
[16, 5120]
blk.60.ssm_conv1d.bias
F32
F32
[5120]
blk.60.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.60.ssm_d
F32
F32
[5120]
blk.60.ssm_dt.bias
F32
F32
[5120]
blk.60.ssm_dt.weight
F16
F16
[160, 5120]
blk.60.ssm_in.weight
F16
F16
[2560, 10240]
blk.60.ssm_out.weight
F16
F16
[5120, 2560]
blk.60.ssm_x.weight
F16
F16
[5120, 192]
blk.61
blk.61.attn_norm.weight
F32
F32
[2560]
blk.61.ssm_a
F32
F32
[16, 5120]
blk.61.ssm_conv1d.bias
F32
F32
[5120]
blk.61.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.61.ssm_d
F32
F32
[5120]
blk.61.ssm_dt.bias
F32
F32
[5120]
blk.61.ssm_dt.weight
F16
F16
[160, 5120]
blk.61.ssm_in.weight
F16
F16
[2560, 10240]
blk.61.ssm_out.weight
F16
F16
[5120, 2560]
blk.61.ssm_x.weight
F16
F16
[5120, 192]
blk.62
blk.62.attn_norm.weight
F32
F32
[2560]
blk.62.ssm_a
F32
F32
[16, 5120]
blk.62.ssm_conv1d.bias
F32
F32
[5120]
blk.62.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.62.ssm_d
F32
F32
[5120]
blk.62.ssm_dt.bias
F32
F32
[5120]
blk.62.ssm_dt.weight
F16
F16
[160, 5120]
blk.62.ssm_in.weight
F16
F16
[2560, 10240]
blk.62.ssm_out.weight
F16
F16
[5120, 2560]
blk.62.ssm_x.weight
F16
F16
[5120, 192]
blk.63
blk.63.attn_norm.weight
F32
F32
[2560]
blk.63.ssm_a
F32
F32
[16, 5120]
blk.63.ssm_conv1d.bias
F32
F32
[5120]
blk.63.ssm_conv1d.weight
F32
F32
[4, 5120]
blk.63.ssm_d
F32
F32
[5120]
blk.63.ssm_dt.bias
F32
F32
[5120]
blk.63.ssm_dt.weight
F16
F16
[160, 5120]
blk.63.ssm_in.weight
F16
F16
[2560, 10240]
blk.63.ssm_out.weight
F16
F16
[5120, 2560]
blk.63.ssm_x.weight
F16
F16
[5120, 192]
output_norm.weight
F32
F32
[2560]