mollysama/rwkv-7-g1:0.1B/model

mollysama/

rwkv-7-g1:0.1B

1,358 Downloads Updated 3 months ago

thinking 0.1b 0.4b 1.5b 2.9b

rwkv-7-g1:0.1B ... /

model

a86b93d4dc1f · 212MB

Metadata

general.architecture

rwkv7

rwkv7
general.file_type

Q8_0

Q8_0
rwkv7.attention.decay_lora_rank

64

64
rwkv7.attention.gate_lora_rank

128

128
rwkv7.attention.head_count

0

0
rwkv7.attention.iclr_lora_rank

64

64
rwkv7.attention.layer_norm_epsilon

1e-05

1e-05
rwkv7.attention.value_residual_mix_lora_rank

32

32
rwkv7.block_count

12

12
rwkv7.context_length

1048576

1048576
rwkv7.embedding_length

768

768
rwkv7.feed_forward_length

3072

3072
rwkv7.wkv.head_size

64

64
tokenizer.ggml.bos_token_id

0

0
tokenizer.ggml.eos_token_id

0

0
tokenizer.ggml.eot_token_id

261

261
tokenizer.ggml.model

rwkv

rwkv
tokenizer.ggml.token_type

[3, 1, 1, 1, 1, ...]

[3, 1, 1, 1, 1, ...]
tokenizer.ggml.tokens

[<s>, \x00, \x01, \x02, \x03, ...]

[<s>, \x00, \x01, \x02, \x03, ...]

Tensor

Name

Type

Shape
token_embd.weight

Q8_0

Q8_0

[768, 65536]

blk.0

blk.0.attn_norm.bias

F32

F32

[768]
blk.0.attn_norm.weight

F32

F32

[768]
blk.0.attn_norm_2.bias

F32

F32

[768]
blk.0.attn_norm_2.weight

F32

F32

[768]
blk.0.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.0.channel_mix_lerp_k.weight

F32

F32

[768]
blk.0.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.0.time_mix_a0.weight

F32

F32

[768]
blk.0.time_mix_a1.weight

F16

F16

[768, 64]
blk.0.time_mix_a2.weight

F16

F16

[64, 768]
blk.0.time_mix_g1.weight

F16

F16

[768, 128]
blk.0.time_mix_g2.weight

F16

F16

[128, 768]
blk.0.time_mix_k_a.weight

F32

F32

[768]
blk.0.time_mix_k_k.weight

F32

F32

[768]
blk.0.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.0.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.0.time_mix_ln.bias

F32

F32

[768]
blk.0.time_mix_ln.weight

F32

F32

[768]
blk.0.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.0.time_mix_r_k.weight

F32

F32

[768]
blk.0.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.0.time_mix_v0.weight

F32

F32

[768]
blk.0.time_mix_v1.weight

F16

F16

[768, 64]
blk.0.time_mix_v2.weight

F16

F16

[64, 768]
blk.0.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.0.time_mix_w0.weight

F32

F32

[768]
blk.0.time_mix_w1.weight

F32

F32

[768, 64]
blk.0.time_mix_w2.weight

F32

F32

[64, 768]

blk.1

blk.1.attn_norm.bias

F32

F32

[768]
blk.1.attn_norm.weight

F32

F32

[768]
blk.1.attn_norm_2.bias

F32

F32

[768]
blk.1.attn_norm_2.weight

F32

F32

[768]
blk.1.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.1.channel_mix_lerp_k.weight

F32

F32

[768]
blk.1.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.1.time_mix_a0.weight

F32

F32

[768]
blk.1.time_mix_a1.weight

F16

F16

[768, 64]
blk.1.time_mix_a2.weight

F16

F16

[64, 768]
blk.1.time_mix_g1.weight

F16

F16

[768, 128]
blk.1.time_mix_g2.weight

F16

F16

[128, 768]
blk.1.time_mix_k_a.weight

F32

F32

[768]
blk.1.time_mix_k_k.weight

F32

F32

[768]
blk.1.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.1.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.1.time_mix_ln.bias

F32

F32

[768]
blk.1.time_mix_ln.weight

F32

F32

[768]
blk.1.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.1.time_mix_r_k.weight

F32

F32

[768]
blk.1.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.1.time_mix_v0.weight

F32

F32

[768]
blk.1.time_mix_v1.weight

F16

F16

[768, 32]
blk.1.time_mix_v2.weight

F16

F16

[32, 768]
blk.1.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.1.time_mix_w0.weight

F32

F32

[768]
blk.1.time_mix_w1.weight

F32

F32

[768, 64]
blk.1.time_mix_w2.weight

F32

F32

[64, 768]

blk.2

blk.2.attn_norm.bias

F32

F32

[768]
blk.2.attn_norm.weight

F32

F32

[768]
blk.2.attn_norm_2.bias

F32

F32

[768]
blk.2.attn_norm_2.weight

F32

F32

[768]
blk.2.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.2.channel_mix_lerp_k.weight

F32

F32

[768]
blk.2.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.2.time_mix_a0.weight

F32

F32

[768]
blk.2.time_mix_a1.weight

F16

F16

[768, 64]
blk.2.time_mix_a2.weight

F16

F16

[64, 768]
blk.2.time_mix_g1.weight

F16

F16

[768, 128]
blk.2.time_mix_g2.weight

F16

F16

[128, 768]
blk.2.time_mix_k_a.weight

F32

F32

[768]
blk.2.time_mix_k_k.weight

F32

F32

[768]
blk.2.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.2.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.2.time_mix_ln.bias

F32

F32

[768]
blk.2.time_mix_ln.weight

F32

F32

[768]
blk.2.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.2.time_mix_r_k.weight

F32

F32

[768]
blk.2.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.2.time_mix_v0.weight

F32

F32

[768]
blk.2.time_mix_v1.weight

F16

F16

[768, 32]
blk.2.time_mix_v2.weight

F16

F16

[32, 768]
blk.2.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.2.time_mix_w0.weight

F32

F32

[768]
blk.2.time_mix_w1.weight

F32

F32

[768, 64]
blk.2.time_mix_w2.weight

F32

F32

[64, 768]

blk.3

blk.3.attn_norm.bias

F32

F32

[768]
blk.3.attn_norm.weight

F32

F32

[768]
blk.3.attn_norm_2.bias

F32

F32

[768]
blk.3.attn_norm_2.weight

F32

F32

[768]
blk.3.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.3.channel_mix_lerp_k.weight

F32

F32

[768]
blk.3.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.3.time_mix_a0.weight

F32

F32

[768]
blk.3.time_mix_a1.weight

F16

F16

[768, 64]
blk.3.time_mix_a2.weight

F16

F16

[64, 768]
blk.3.time_mix_g1.weight

F16

F16

[768, 128]
blk.3.time_mix_g2.weight

F16

F16

[128, 768]
blk.3.time_mix_k_a.weight

F32

F32

[768]
blk.3.time_mix_k_k.weight

F32

F32

[768]
blk.3.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.3.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.3.time_mix_ln.bias

F32

F32

[768]
blk.3.time_mix_ln.weight

F32

F32

[768]
blk.3.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.3.time_mix_r_k.weight

F32

F32

[768]
blk.3.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.3.time_mix_v0.weight

F32

F32

[768]
blk.3.time_mix_v1.weight

F16

F16

[768, 32]
blk.3.time_mix_v2.weight

F16

F16

[32, 768]
blk.3.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.3.time_mix_w0.weight

F32

F32

[768]
blk.3.time_mix_w1.weight

F32

F32

[768, 64]
blk.3.time_mix_w2.weight

F32

F32

[64, 768]

blk.4

blk.4.attn_norm.bias

F32

F32

[768]
blk.4.attn_norm.weight

F32

F32

[768]
blk.4.attn_norm_2.bias

F32

F32

[768]
blk.4.attn_norm_2.weight

F32

F32

[768]
blk.4.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.4.channel_mix_lerp_k.weight

F32

F32

[768]
blk.4.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.4.time_mix_a0.weight

F32

F32

[768]
blk.4.time_mix_a1.weight

F16

F16

[768, 64]
blk.4.time_mix_a2.weight

F16

F16

[64, 768]
blk.4.time_mix_g1.weight

F16

F16

[768, 128]
blk.4.time_mix_g2.weight

F16

F16

[128, 768]
blk.4.time_mix_k_a.weight

F32

F32

[768]
blk.4.time_mix_k_k.weight

F32

F32

[768]
blk.4.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.4.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.4.time_mix_ln.bias

F32

F32

[768]
blk.4.time_mix_ln.weight

F32

F32

[768]
blk.4.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.4.time_mix_r_k.weight

F32

F32

[768]
blk.4.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.4.time_mix_v0.weight

F32

F32

[768]
blk.4.time_mix_v1.weight

F16

F16

[768, 32]
blk.4.time_mix_v2.weight

F16

F16

[32, 768]
blk.4.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.4.time_mix_w0.weight

F32

F32

[768]
blk.4.time_mix_w1.weight

F32

F32

[768, 64]
blk.4.time_mix_w2.weight

F32

F32

[64, 768]

blk.5

blk.5.attn_norm.bias

F32

F32

[768]
blk.5.attn_norm.weight

F32

F32

[768]
blk.5.attn_norm_2.bias

F32

F32

[768]
blk.5.attn_norm_2.weight

F32

F32

[768]
blk.5.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.5.channel_mix_lerp_k.weight

F32

F32

[768]
blk.5.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.5.time_mix_a0.weight

F32

F32

[768]
blk.5.time_mix_a1.weight

F16

F16

[768, 64]
blk.5.time_mix_a2.weight

F16

F16

[64, 768]
blk.5.time_mix_g1.weight

F16

F16

[768, 128]
blk.5.time_mix_g2.weight

F16

F16

[128, 768]
blk.5.time_mix_k_a.weight

F32

F32

[768]
blk.5.time_mix_k_k.weight

F32

F32

[768]
blk.5.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.5.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.5.time_mix_ln.bias

F32

F32

[768]
blk.5.time_mix_ln.weight

F32

F32

[768]
blk.5.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.5.time_mix_r_k.weight

F32

F32

[768]
blk.5.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.5.time_mix_v0.weight

F32

F32

[768]
blk.5.time_mix_v1.weight

F16

F16

[768, 32]
blk.5.time_mix_v2.weight

F16

F16

[32, 768]
blk.5.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.5.time_mix_w0.weight

F32

F32

[768]
blk.5.time_mix_w1.weight

F32

F32

[768, 64]
blk.5.time_mix_w2.weight

F32

F32

[64, 768]

blk.6

blk.6.attn_norm.bias

F32

F32

[768]
blk.6.attn_norm.weight

F32

F32

[768]
blk.6.attn_norm_2.bias

F32

F32

[768]
blk.6.attn_norm_2.weight

F32

F32

[768]
blk.6.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.6.channel_mix_lerp_k.weight

F32

F32

[768]
blk.6.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.6.time_mix_a0.weight

F32

F32

[768]
blk.6.time_mix_a1.weight

F16

F16

[768, 64]
blk.6.time_mix_a2.weight

F16

F16

[64, 768]
blk.6.time_mix_g1.weight

F16

F16

[768, 128]
blk.6.time_mix_g2.weight

F16

F16

[128, 768]
blk.6.time_mix_k_a.weight

F32

F32

[768]
blk.6.time_mix_k_k.weight

F32

F32

[768]
blk.6.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.6.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.6.time_mix_ln.bias

F32

F32

[768]
blk.6.time_mix_ln.weight

F32

F32

[768]
blk.6.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.6.time_mix_r_k.weight

F32

F32

[768]
blk.6.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.6.time_mix_v0.weight

F32

F32

[768]
blk.6.time_mix_v1.weight

F16

F16

[768, 32]
blk.6.time_mix_v2.weight

F16

F16

[32, 768]
blk.6.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.6.time_mix_w0.weight

F32

F32

[768]
blk.6.time_mix_w1.weight

F32

F32

[768, 64]
blk.6.time_mix_w2.weight

F32

F32

[64, 768]

blk.7

blk.7.attn_norm.bias

F32

F32

[768]
blk.7.attn_norm.weight

F32

F32

[768]
blk.7.attn_norm_2.bias

F32

F32

[768]
blk.7.attn_norm_2.weight

F32

F32

[768]
blk.7.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.7.channel_mix_lerp_k.weight

F32

F32

[768]
blk.7.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.7.time_mix_a0.weight

F32

F32

[768]
blk.7.time_mix_a1.weight

F16

F16

[768, 64]
blk.7.time_mix_a2.weight

F16

F16

[64, 768]
blk.7.time_mix_g1.weight

F16

F16

[768, 128]
blk.7.time_mix_g2.weight

F16

F16

[128, 768]
blk.7.time_mix_k_a.weight

F32

F32

[768]
blk.7.time_mix_k_k.weight

F32

F32

[768]
blk.7.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.7.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.7.time_mix_ln.bias

F32

F32

[768]
blk.7.time_mix_ln.weight

F32

F32

[768]
blk.7.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.7.time_mix_r_k.weight

F32

F32

[768]
blk.7.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.7.time_mix_v0.weight

F32

F32

[768]
blk.7.time_mix_v1.weight

F16

F16

[768, 32]
blk.7.time_mix_v2.weight

F16

F16

[32, 768]
blk.7.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.7.time_mix_w0.weight

F32

F32

[768]
blk.7.time_mix_w1.weight

F32

F32

[768, 64]
blk.7.time_mix_w2.weight

F32

F32

[64, 768]

blk.8

blk.8.attn_norm.bias

F32

F32

[768]
blk.8.attn_norm.weight

F32

F32

[768]
blk.8.attn_norm_2.bias

F32

F32

[768]
blk.8.attn_norm_2.weight

F32

F32

[768]
blk.8.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.8.channel_mix_lerp_k.weight

F32

F32

[768]
blk.8.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.8.time_mix_a0.weight

F32

F32

[768]
blk.8.time_mix_a1.weight

F16

F16

[768, 64]
blk.8.time_mix_a2.weight

F16

F16

[64, 768]
blk.8.time_mix_g1.weight

F16

F16

[768, 128]
blk.8.time_mix_g2.weight

F16

F16

[128, 768]
blk.8.time_mix_k_a.weight

F32

F32

[768]
blk.8.time_mix_k_k.weight

F32

F32

[768]
blk.8.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.8.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.8.time_mix_ln.bias

F32

F32

[768]
blk.8.time_mix_ln.weight

F32

F32

[768]
blk.8.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.8.time_mix_r_k.weight

F32

F32

[768]
blk.8.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.8.time_mix_v0.weight

F32

F32

[768]
blk.8.time_mix_v1.weight

F16

F16

[768, 32]
blk.8.time_mix_v2.weight

F16

F16

[32, 768]
blk.8.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.8.time_mix_w0.weight

F32

F32

[768]
blk.8.time_mix_w1.weight

F32

F32

[768, 64]
blk.8.time_mix_w2.weight

F32

F32

[64, 768]

blk.9

blk.9.attn_norm.bias

F32

F32

[768]
blk.9.attn_norm.weight

F32

F32

[768]
blk.9.attn_norm_2.bias

F32

F32

[768]
blk.9.attn_norm_2.weight

F32

F32

[768]
blk.9.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.9.channel_mix_lerp_k.weight

F32

F32

[768]
blk.9.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.9.time_mix_a0.weight

F32

F32

[768]
blk.9.time_mix_a1.weight

F16

F16

[768, 64]
blk.9.time_mix_a2.weight

F16

F16

[64, 768]
blk.9.time_mix_g1.weight

F16

F16

[768, 128]
blk.9.time_mix_g2.weight

F16

F16

[128, 768]
blk.9.time_mix_k_a.weight

F32

F32

[768]
blk.9.time_mix_k_k.weight

F32

F32

[768]
blk.9.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.9.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.9.time_mix_ln.bias

F32

F32

[768]
blk.9.time_mix_ln.weight

F32

F32

[768]
blk.9.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.9.time_mix_r_k.weight

F32

F32

[768]
blk.9.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.9.time_mix_v0.weight

F32

F32

[768]
blk.9.time_mix_v1.weight

F16

F16

[768, 32]
blk.9.time_mix_v2.weight

F16

F16

[32, 768]
blk.9.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.9.time_mix_w0.weight

F32

F32

[768]
blk.9.time_mix_w1.weight

F32

F32

[768, 64]
blk.9.time_mix_w2.weight

F32

F32

[64, 768]

blk.10

blk.10.attn_norm.bias

F32

F32

[768]
blk.10.attn_norm.weight

F32

F32

[768]
blk.10.attn_norm_2.bias

F32

F32

[768]
blk.10.attn_norm_2.weight

F32

F32

[768]
blk.10.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.10.channel_mix_lerp_k.weight

F32

F32

[768]
blk.10.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.10.time_mix_a0.weight

F32

F32

[768]
blk.10.time_mix_a1.weight

F16

F16

[768, 64]
blk.10.time_mix_a2.weight

F16

F16

[64, 768]
blk.10.time_mix_g1.weight

F16

F16

[768, 128]
blk.10.time_mix_g2.weight

F16

F16

[128, 768]
blk.10.time_mix_k_a.weight

F32

F32

[768]
blk.10.time_mix_k_k.weight

F32

F32

[768]
blk.10.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.10.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.10.time_mix_ln.bias

F32

F32

[768]
blk.10.time_mix_ln.weight

F32

F32

[768]
blk.10.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.10.time_mix_r_k.weight

F32

F32

[768]
blk.10.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.10.time_mix_v0.weight

F32

F32

[768]
blk.10.time_mix_v1.weight

F16

F16

[768, 32]
blk.10.time_mix_v2.weight

F16

F16

[32, 768]
blk.10.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.10.time_mix_w0.weight

F32

F32

[768]
blk.10.time_mix_w1.weight

F32

F32

[768, 64]
blk.10.time_mix_w2.weight

F32

F32

[64, 768]

blk.11

blk.11.attn_norm.bias

F32

F32

[768]
blk.11.attn_norm.weight

F32

F32

[768]
blk.11.attn_norm_2.bias

F32

F32

[768]
blk.11.attn_norm_2.weight

F32

F32

[768]
blk.11.channel_mix_key.weight

Q8_0

Q8_0

[768, 3072]
blk.11.channel_mix_lerp_k.weight

F32

F32

[768]
blk.11.channel_mix_value.weight

Q8_0

Q8_0

[3072, 768]
blk.11.time_mix_a0.weight

F32

F32

[768]
blk.11.time_mix_a1.weight

F16

F16

[768, 64]
blk.11.time_mix_a2.weight

F16

F16

[64, 768]
blk.11.time_mix_g1.weight

F16

F16

[768, 128]
blk.11.time_mix_g2.weight

F16

F16

[128, 768]
blk.11.time_mix_k_a.weight

F32

F32

[768]
blk.11.time_mix_k_k.weight

F32

F32

[768]
blk.11.time_mix_key.weight

Q8_0

Q8_0

[768, 768]
blk.11.time_mix_lerp_fused.weight

F32

F32

[768, 1, 1, 6]
blk.11.time_mix_ln.bias

F32

F32

[768]
blk.11.time_mix_ln.weight

F32

F32

[768]
blk.11.time_mix_output.weight

Q8_0

Q8_0

[768, 768]
blk.11.time_mix_r_k.weight

F32

F32

[768]
blk.11.time_mix_receptance.weight

Q8_0

Q8_0

[768, 768]
blk.11.time_mix_v0.weight

F32

F32

[768]
blk.11.time_mix_v1.weight

F16

F16

[768, 32]
blk.11.time_mix_v2.weight

F16

F16

[32, 768]
blk.11.time_mix_value.weight

Q8_0

Q8_0

[768, 768]
blk.11.time_mix_w0.weight

F32

F32

[768]
blk.11.time_mix_w1.weight

F32

F32

[768, 64]
blk.11.time_mix_w2.weight

F32

F32

[64, 768]

output.weight

Q8_0

Q8_0

[768, 65536]
output_norm.bias

F32

F32

[768]
token_embd_norm.bias

F32

F32

[768]
token_embd_norm.weight

F32

F32

[768]
output_norm.weight

F32

F32

[768]