GitHub
186.4K
Download
Models
Docs
Pricing
Sign in
Download
Models
Docs
Pricing
Sign in
ripple-preview
:65b-gguf-bf16
32
Downloads
Updated
18 minutes ago
65B dense model built for agentic coding, with up to 1M tokens of context.
65B dense model built for agentic coding, with up to 1M tokens of context.
Cancel
tools
thinking
65b
ripple-preview:65b-gguf-bf16
...
/
model
5ef95c7c3782 · 130GB
Metadata
general.architecture
ricursive
ricursive
general.file_type
BF16
BF16
ricursive.attention.head_count
80
80
ricursive.attention.head_count_kv
8
8
ricursive.attention.key_length
128
128
ricursive.attention.layer_norm_rms_epsilon
1e-06
1e-06
ricursive.attention.rope_pattern
[true, true, true, true, false, ...]
[true, true, true, true, false, ...]
ricursive.attention.sliding_window
1024
1024
ricursive.attention.sliding_window_pattern
[true, true, true, true, false, ...]
[true, true, true, true, false, ...]
ricursive.attention.value_length
128
128
ricursive.block_count
80
80
ricursive.context_length
2097152
2097152
ricursive.embedding_length
5120
5120
ricursive.embedding_scale
71.55418
71.55418
ricursive.feed_forward_length
40960
40960
ricursive.rope.dimension_count
128
128
ricursive.rope.freq_base
10000
10000
ricursive.swiglu_clamp
[20, 20, 20, 20, 20, ...]
[20, 20, 20, 20, 20, ...]
ricursive.vocab_size
100352
100352
tokenizer.ggml.bos_token_id
100257
100257
tokenizer.ggml.clean_spaces
false
false
tokenizer.ggml.eos_token_id
100265
100265
tokenizer.ggml.eot_token_id
100265
100265
tokenizer.ggml.merges
[Ġ Ġ, ĠĠ ĠĠ, i n, Ġ t, ĠĠĠĠ ĠĠĠĠ, ...]
[Ġ Ġ, ĠĠ ĠĠ, i n, Ġ t, ĠĠĠĠ ĠĠĠĠ, ...]
tokenizer.ggml.model
gpt2
gpt2
tokenizer.ggml.padding_token_id
100277
100277
tokenizer.ggml.pre
dbrx
dbrx
tokenizer.ggml.token_type
[1, 1, 1, 1, 1, ...]
[1, 1, 1, 1, 1, ...]
tokenizer.ggml.tokens
[!, ", #, $, %, ...]
[!, ", #, $, %, ...]
tokenizer.ggml.unknown_token_id
100257
100257
Tensor
Name
Type
Shape
token_embd.weight
BF16
BF16
[5120, 100352]
blk.0
blk.0.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.0.attn_k.weight
BF16
BF16
[5120, 1024]
blk.0.attn_k_norm.weight
F32
F32
[128]
blk.0.attn_norm.weight
F32
F32
[5120]
blk.0.attn_output.weight
BF16
BF16
[10240, 5120]
blk.0.attn_q.weight
BF16
BF16
[5120, 10240]
blk.0.attn_q_norm.weight
F32
F32
[128]
blk.0.attn_v.weight
BF16
BF16
[5120, 1024]
blk.0.attn_vres_lambda.weight
F32
F32
[1]
blk.0.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.0.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.0.ffn_norm.weight
F32
F32
[5120]
blk.0.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.0.post_attention_norm.weight
F32
F32
[5120]
blk.0.post_ffw_norm.weight
F32
F32
[5120]
blk.1
blk.1.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.1.attn_k.weight
BF16
BF16
[5120, 1024]
blk.1.attn_k_norm.weight
F32
F32
[128]
blk.1.attn_norm.weight
F32
F32
[5120]
blk.1.attn_output.weight
BF16
BF16
[10240, 5120]
blk.1.attn_q.weight
BF16
BF16
[5120, 10240]
blk.1.attn_q_norm.weight
F32
F32
[128]
blk.1.attn_v.weight
BF16
BF16
[5120, 1024]
blk.1.attn_vres_lambda.weight
F32
F32
[1]
blk.1.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.1.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.1.ffn_norm.weight
F32
F32
[5120]
blk.1.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.1.post_attention_norm.weight
F32
F32
[5120]
blk.1.post_ffw_norm.weight
F32
F32
[5120]
blk.2
blk.2.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.2.attn_k.weight
BF16
BF16
[5120, 1024]
blk.2.attn_k_norm.weight
F32
F32
[128]
blk.2.attn_norm.weight
F32
F32
[5120]
blk.2.attn_output.weight
BF16
BF16
[10240, 5120]
blk.2.attn_q.weight
BF16
BF16
[5120, 10240]
blk.2.attn_q_norm.weight
F32
F32
[128]
blk.2.attn_v.weight
BF16
BF16
[5120, 1024]
blk.2.attn_vres_lambda.weight
F32
F32
[1]
blk.2.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.2.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.2.ffn_norm.weight
F32
F32
[5120]
blk.2.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.2.post_attention_norm.weight
F32
F32
[5120]
blk.2.post_ffw_norm.weight
F32
F32
[5120]
blk.3
blk.3.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.3.attn_k.weight
BF16
BF16
[5120, 1024]
blk.3.attn_k_norm.weight
F32
F32
[128]
blk.3.attn_norm.weight
F32
F32
[5120]
blk.3.attn_output.weight
BF16
BF16
[10240, 5120]
blk.3.attn_q.weight
BF16
BF16
[5120, 10240]
blk.3.attn_q_norm.weight
F32
F32
[128]
blk.3.attn_v.weight
BF16
BF16
[5120, 1024]
blk.3.attn_vres_lambda.weight
F32
F32
[1]
blk.3.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.3.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.3.ffn_norm.weight
F32
F32
[5120]
blk.3.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.3.post_attention_norm.weight
F32
F32
[5120]
blk.3.post_ffw_norm.weight
F32
F32
[5120]
blk.4
blk.4.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.4.attn_k.weight
BF16
BF16
[5120, 1024]
blk.4.attn_k_norm.weight
F32
F32
[128]
blk.4.attn_norm.weight
F32
F32
[5120]
blk.4.attn_output.weight
BF16
BF16
[10240, 5120]
blk.4.attn_q.weight
BF16
BF16
[5120, 10240]
blk.4.attn_q_norm.weight
F32
F32
[128]
blk.4.attn_v.weight
BF16
BF16
[5120, 1024]
blk.4.attn_vres_lambda.weight
F32
F32
[1]
blk.4.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.4.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.4.ffn_norm.weight
F32
F32
[5120]
blk.4.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.4.post_attention_norm.weight
F32
F32
[5120]
blk.4.post_ffw_norm.weight
F32
F32
[5120]
blk.5
blk.5.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.5.attn_k.weight
BF16
BF16
[5120, 1024]
blk.5.attn_k_norm.weight
F32
F32
[128]
blk.5.attn_norm.weight
F32
F32
[5120]
blk.5.attn_output.weight
BF16
BF16
[10240, 5120]
blk.5.attn_q.weight
BF16
BF16
[5120, 10240]
blk.5.attn_q_norm.weight
F32
F32
[128]
blk.5.attn_v.weight
BF16
BF16
[5120, 1024]
blk.5.attn_vres_lambda.weight
F32
F32
[1]
blk.5.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.5.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.5.ffn_norm.weight
F32
F32
[5120]
blk.5.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.5.post_attention_norm.weight
F32
F32
[5120]
blk.5.post_ffw_norm.weight
F32
F32
[5120]
blk.6
blk.6.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.6.attn_k.weight
BF16
BF16
[5120, 1024]
blk.6.attn_k_norm.weight
F32
F32
[128]
blk.6.attn_norm.weight
F32
F32
[5120]
blk.6.attn_output.weight
BF16
BF16
[10240, 5120]
blk.6.attn_q.weight
BF16
BF16
[5120, 10240]
blk.6.attn_q_norm.weight
F32
F32
[128]
blk.6.attn_v.weight
BF16
BF16
[5120, 1024]
blk.6.attn_vres_lambda.weight
F32
F32
[1]
blk.6.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.6.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.6.ffn_norm.weight
F32
F32
[5120]
blk.6.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.6.post_attention_norm.weight
F32
F32
[5120]
blk.6.post_ffw_norm.weight
F32
F32
[5120]
blk.7
blk.7.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.7.attn_k.weight
BF16
BF16
[5120, 1024]
blk.7.attn_k_norm.weight
F32
F32
[128]
blk.7.attn_norm.weight
F32
F32
[5120]
blk.7.attn_output.weight
BF16
BF16
[10240, 5120]
blk.7.attn_q.weight
BF16
BF16
[5120, 10240]
blk.7.attn_q_norm.weight
F32
F32
[128]
blk.7.attn_v.weight
BF16
BF16
[5120, 1024]
blk.7.attn_vres_lambda.weight
F32
F32
[1]
blk.7.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.7.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.7.ffn_norm.weight
F32
F32
[5120]
blk.7.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.7.post_attention_norm.weight
F32
F32
[5120]
blk.7.post_ffw_norm.weight
F32
F32
[5120]
blk.8
blk.8.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.8.attn_k.weight
BF16
BF16
[5120, 1024]
blk.8.attn_k_norm.weight
F32
F32
[128]
blk.8.attn_norm.weight
F32
F32
[5120]
blk.8.attn_output.weight
BF16
BF16
[10240, 5120]
blk.8.attn_q.weight
BF16
BF16
[5120, 10240]
blk.8.attn_q_norm.weight
F32
F32
[128]
blk.8.attn_v.weight
BF16
BF16
[5120, 1024]
blk.8.attn_vres_lambda.weight
F32
F32
[1]
blk.8.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.8.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.8.ffn_norm.weight
F32
F32
[5120]
blk.8.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.8.post_attention_norm.weight
F32
F32
[5120]
blk.8.post_ffw_norm.weight
F32
F32
[5120]
blk.9
blk.9.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.9.attn_k.weight
BF16
BF16
[5120, 1024]
blk.9.attn_k_norm.weight
F32
F32
[128]
blk.9.attn_norm.weight
F32
F32
[5120]
blk.9.attn_output.weight
BF16
BF16
[10240, 5120]
blk.9.attn_q.weight
BF16
BF16
[5120, 10240]
blk.9.attn_q_norm.weight
F32
F32
[128]
blk.9.attn_v.weight
BF16
BF16
[5120, 1024]
blk.9.attn_vres_lambda.weight
F32
F32
[1]
blk.9.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.9.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.9.ffn_norm.weight
F32
F32
[5120]
blk.9.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.9.post_attention_norm.weight
F32
F32
[5120]
blk.9.post_ffw_norm.weight
F32
F32
[5120]
blk.10
blk.10.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.10.attn_k.weight
BF16
BF16
[5120, 1024]
blk.10.attn_k_norm.weight
F32
F32
[128]
blk.10.attn_norm.weight
F32
F32
[5120]
blk.10.attn_output.weight
BF16
BF16
[10240, 5120]
blk.10.attn_q.weight
BF16
BF16
[5120, 10240]
blk.10.attn_q_norm.weight
F32
F32
[128]
blk.10.attn_v.weight
BF16
BF16
[5120, 1024]
blk.10.attn_vres_lambda.weight
F32
F32
[1]
blk.10.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.10.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.10.ffn_norm.weight
F32
F32
[5120]
blk.10.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.10.post_attention_norm.weight
F32
F32
[5120]
blk.10.post_ffw_norm.weight
F32
F32
[5120]
blk.11
blk.11.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.11.attn_k.weight
BF16
BF16
[5120, 1024]
blk.11.attn_k_norm.weight
F32
F32
[128]
blk.11.attn_norm.weight
F32
F32
[5120]
blk.11.attn_output.weight
BF16
BF16
[10240, 5120]
blk.11.attn_q.weight
BF16
BF16
[5120, 10240]
blk.11.attn_q_norm.weight
F32
F32
[128]
blk.11.attn_v.weight
BF16
BF16
[5120, 1024]
blk.11.attn_vres_lambda.weight
F32
F32
[1]
blk.11.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.11.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.11.ffn_norm.weight
F32
F32
[5120]
blk.11.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.11.post_attention_norm.weight
F32
F32
[5120]
blk.11.post_ffw_norm.weight
F32
F32
[5120]
blk.12
blk.12.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.12.attn_k.weight
BF16
BF16
[5120, 1024]
blk.12.attn_k_norm.weight
F32
F32
[128]
blk.12.attn_norm.weight
F32
F32
[5120]
blk.12.attn_output.weight
BF16
BF16
[10240, 5120]
blk.12.attn_q.weight
BF16
BF16
[5120, 10240]
blk.12.attn_q_norm.weight
F32
F32
[128]
blk.12.attn_v.weight
BF16
BF16
[5120, 1024]
blk.12.attn_vres_lambda.weight
F32
F32
[1]
blk.12.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.12.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.12.ffn_norm.weight
F32
F32
[5120]
blk.12.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.12.post_attention_norm.weight
F32
F32
[5120]
blk.12.post_ffw_norm.weight
F32
F32
[5120]
blk.13
blk.13.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.13.attn_k.weight
BF16
BF16
[5120, 1024]
blk.13.attn_k_norm.weight
F32
F32
[128]
blk.13.attn_norm.weight
F32
F32
[5120]
blk.13.attn_output.weight
BF16
BF16
[10240, 5120]
blk.13.attn_q.weight
BF16
BF16
[5120, 10240]
blk.13.attn_q_norm.weight
F32
F32
[128]
blk.13.attn_v.weight
BF16
BF16
[5120, 1024]
blk.13.attn_vres_lambda.weight
F32
F32
[1]
blk.13.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.13.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.13.ffn_norm.weight
F32
F32
[5120]
blk.13.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.13.post_attention_norm.weight
F32
F32
[5120]
blk.13.post_ffw_norm.weight
F32
F32
[5120]
blk.14
blk.14.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.14.attn_k.weight
BF16
BF16
[5120, 1024]
blk.14.attn_k_norm.weight
F32
F32
[128]
blk.14.attn_norm.weight
F32
F32
[5120]
blk.14.attn_output.weight
BF16
BF16
[10240, 5120]
blk.14.attn_q.weight
BF16
BF16
[5120, 10240]
blk.14.attn_q_norm.weight
F32
F32
[128]
blk.14.attn_v.weight
BF16
BF16
[5120, 1024]
blk.14.attn_vres_lambda.weight
F32
F32
[1]
blk.14.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.14.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.14.ffn_norm.weight
F32
F32
[5120]
blk.14.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.14.post_attention_norm.weight
F32
F32
[5120]
blk.14.post_ffw_norm.weight
F32
F32
[5120]
blk.15
blk.15.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.15.attn_k.weight
BF16
BF16
[5120, 1024]
blk.15.attn_k_norm.weight
F32
F32
[128]
blk.15.attn_norm.weight
F32
F32
[5120]
blk.15.attn_output.weight
BF16
BF16
[10240, 5120]
blk.15.attn_q.weight
BF16
BF16
[5120, 10240]
blk.15.attn_q_norm.weight
F32
F32
[128]
blk.15.attn_v.weight
BF16
BF16
[5120, 1024]
blk.15.attn_vres_lambda.weight
F32
F32
[1]
blk.15.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.15.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.15.ffn_norm.weight
F32
F32
[5120]
blk.15.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.15.post_attention_norm.weight
F32
F32
[5120]
blk.15.post_ffw_norm.weight
F32
F32
[5120]
blk.16
blk.16.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.16.attn_k.weight
BF16
BF16
[5120, 1024]
blk.16.attn_k_norm.weight
F32
F32
[128]
blk.16.attn_norm.weight
F32
F32
[5120]
blk.16.attn_output.weight
BF16
BF16
[10240, 5120]
blk.16.attn_q.weight
BF16
BF16
[5120, 10240]
blk.16.attn_q_norm.weight
F32
F32
[128]
blk.16.attn_v.weight
BF16
BF16
[5120, 1024]
blk.16.attn_vres_lambda.weight
F32
F32
[1]
blk.16.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.16.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.16.ffn_norm.weight
F32
F32
[5120]
blk.16.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.16.post_attention_norm.weight
F32
F32
[5120]
blk.16.post_ffw_norm.weight
F32
F32
[5120]
blk.17
blk.17.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.17.attn_k.weight
BF16
BF16
[5120, 1024]
blk.17.attn_k_norm.weight
F32
F32
[128]
blk.17.attn_norm.weight
F32
F32
[5120]
blk.17.attn_output.weight
BF16
BF16
[10240, 5120]
blk.17.attn_q.weight
BF16
BF16
[5120, 10240]
blk.17.attn_q_norm.weight
F32
F32
[128]
blk.17.attn_v.weight
BF16
BF16
[5120, 1024]
blk.17.attn_vres_lambda.weight
F32
F32
[1]
blk.17.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.17.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.17.ffn_norm.weight
F32
F32
[5120]
blk.17.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.17.post_attention_norm.weight
F32
F32
[5120]
blk.17.post_ffw_norm.weight
F32
F32
[5120]
blk.18
blk.18.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.18.attn_k.weight
BF16
BF16
[5120, 1024]
blk.18.attn_k_norm.weight
F32
F32
[128]
blk.18.attn_norm.weight
F32
F32
[5120]
blk.18.attn_output.weight
BF16
BF16
[10240, 5120]
blk.18.attn_q.weight
BF16
BF16
[5120, 10240]
blk.18.attn_q_norm.weight
F32
F32
[128]
blk.18.attn_v.weight
BF16
BF16
[5120, 1024]
blk.18.attn_vres_lambda.weight
F32
F32
[1]
blk.18.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.18.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.18.ffn_norm.weight
F32
F32
[5120]
blk.18.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.18.post_attention_norm.weight
F32
F32
[5120]
blk.18.post_ffw_norm.weight
F32
F32
[5120]
blk.19
blk.19.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.19.attn_k.weight
BF16
BF16
[5120, 1024]
blk.19.attn_k_norm.weight
F32
F32
[128]
blk.19.attn_norm.weight
F32
F32
[5120]
blk.19.attn_output.weight
BF16
BF16
[10240, 5120]
blk.19.attn_q.weight
BF16
BF16
[5120, 10240]
blk.19.attn_q_norm.weight
F32
F32
[128]
blk.19.attn_v.weight
BF16
BF16
[5120, 1024]
blk.19.attn_vres_lambda.weight
F32
F32
[1]
blk.19.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.19.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.19.ffn_norm.weight
F32
F32
[5120]
blk.19.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.19.post_attention_norm.weight
F32
F32
[5120]
blk.19.post_ffw_norm.weight
F32
F32
[5120]
blk.20
blk.20.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.20.attn_k.weight
BF16
BF16
[5120, 1024]
blk.20.attn_k_norm.weight
F32
F32
[128]
blk.20.attn_norm.weight
F32
F32
[5120]
blk.20.attn_output.weight
BF16
BF16
[10240, 5120]
blk.20.attn_q.weight
BF16
BF16
[5120, 10240]
blk.20.attn_q_norm.weight
F32
F32
[128]
blk.20.attn_v.weight
BF16
BF16
[5120, 1024]
blk.20.attn_vres_lambda.weight
F32
F32
[1]
blk.20.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.20.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.20.ffn_norm.weight
F32
F32
[5120]
blk.20.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.20.post_attention_norm.weight
F32
F32
[5120]
blk.20.post_ffw_norm.weight
F32
F32
[5120]
blk.21
blk.21.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.21.attn_k.weight
BF16
BF16
[5120, 1024]
blk.21.attn_k_norm.weight
F32
F32
[128]
blk.21.attn_norm.weight
F32
F32
[5120]
blk.21.attn_output.weight
BF16
BF16
[10240, 5120]
blk.21.attn_q.weight
BF16
BF16
[5120, 10240]
blk.21.attn_q_norm.weight
F32
F32
[128]
blk.21.attn_v.weight
BF16
BF16
[5120, 1024]
blk.21.attn_vres_lambda.weight
F32
F32
[1]
blk.21.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.21.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.21.ffn_norm.weight
F32
F32
[5120]
blk.21.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.21.post_attention_norm.weight
F32
F32
[5120]
blk.21.post_ffw_norm.weight
F32
F32
[5120]
blk.22
blk.22.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.22.attn_k.weight
BF16
BF16
[5120, 1024]
blk.22.attn_k_norm.weight
F32
F32
[128]
blk.22.attn_norm.weight
F32
F32
[5120]
blk.22.attn_output.weight
BF16
BF16
[10240, 5120]
blk.22.attn_q.weight
BF16
BF16
[5120, 10240]
blk.22.attn_q_norm.weight
F32
F32
[128]
blk.22.attn_v.weight
BF16
BF16
[5120, 1024]
blk.22.attn_vres_lambda.weight
F32
F32
[1]
blk.22.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.22.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.22.ffn_norm.weight
F32
F32
[5120]
blk.22.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.22.post_attention_norm.weight
F32
F32
[5120]
blk.22.post_ffw_norm.weight
F32
F32
[5120]
blk.23
blk.23.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.23.attn_k.weight
BF16
BF16
[5120, 1024]
blk.23.attn_k_norm.weight
F32
F32
[128]
blk.23.attn_norm.weight
F32
F32
[5120]
blk.23.attn_output.weight
BF16
BF16
[10240, 5120]
blk.23.attn_q.weight
BF16
BF16
[5120, 10240]
blk.23.attn_q_norm.weight
F32
F32
[128]
blk.23.attn_v.weight
BF16
BF16
[5120, 1024]
blk.23.attn_vres_lambda.weight
F32
F32
[1]
blk.23.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.23.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.23.ffn_norm.weight
F32
F32
[5120]
blk.23.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.23.post_attention_norm.weight
F32
F32
[5120]
blk.23.post_ffw_norm.weight
F32
F32
[5120]
blk.24
blk.24.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.24.attn_k.weight
BF16
BF16
[5120, 1024]
blk.24.attn_k_norm.weight
F32
F32
[128]
blk.24.attn_norm.weight
F32
F32
[5120]
blk.24.attn_output.weight
BF16
BF16
[10240, 5120]
blk.24.attn_q.weight
BF16
BF16
[5120, 10240]
blk.24.attn_q_norm.weight
F32
F32
[128]
blk.24.attn_v.weight
BF16
BF16
[5120, 1024]
blk.24.attn_vres_lambda.weight
F32
F32
[1]
blk.24.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.24.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.24.ffn_norm.weight
F32
F32
[5120]
blk.24.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.24.post_attention_norm.weight
F32
F32
[5120]
blk.24.post_ffw_norm.weight
F32
F32
[5120]
blk.25
blk.25.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.25.attn_k.weight
BF16
BF16
[5120, 1024]
blk.25.attn_k_norm.weight
F32
F32
[128]
blk.25.attn_norm.weight
F32
F32
[5120]
blk.25.attn_output.weight
BF16
BF16
[10240, 5120]
blk.25.attn_q.weight
BF16
BF16
[5120, 10240]
blk.25.attn_q_norm.weight
F32
F32
[128]
blk.25.attn_v.weight
BF16
BF16
[5120, 1024]
blk.25.attn_vres_lambda.weight
F32
F32
[1]
blk.25.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.25.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.25.ffn_norm.weight
F32
F32
[5120]
blk.25.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.25.post_attention_norm.weight
F32
F32
[5120]
blk.25.post_ffw_norm.weight
F32
F32
[5120]
blk.26
blk.26.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.26.attn_k.weight
BF16
BF16
[5120, 1024]
blk.26.attn_k_norm.weight
F32
F32
[128]
blk.26.attn_norm.weight
F32
F32
[5120]
blk.26.attn_output.weight
BF16
BF16
[10240, 5120]
blk.26.attn_q.weight
BF16
BF16
[5120, 10240]
blk.26.attn_q_norm.weight
F32
F32
[128]
blk.26.attn_v.weight
BF16
BF16
[5120, 1024]
blk.26.attn_vres_lambda.weight
F32
F32
[1]
blk.26.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.26.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.26.ffn_norm.weight
F32
F32
[5120]
blk.26.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.26.post_attention_norm.weight
F32
F32
[5120]
blk.26.post_ffw_norm.weight
F32
F32
[5120]
blk.27
blk.27.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.27.attn_k.weight
BF16
BF16
[5120, 1024]
blk.27.attn_k_norm.weight
F32
F32
[128]
blk.27.attn_norm.weight
F32
F32
[5120]
blk.27.attn_output.weight
BF16
BF16
[10240, 5120]
blk.27.attn_q.weight
BF16
BF16
[5120, 10240]
blk.27.attn_q_norm.weight
F32
F32
[128]
blk.27.attn_v.weight
BF16
BF16
[5120, 1024]
blk.27.attn_vres_lambda.weight
F32
F32
[1]
blk.27.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.27.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.27.ffn_norm.weight
F32
F32
[5120]
blk.27.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.27.post_attention_norm.weight
F32
F32
[5120]
blk.27.post_ffw_norm.weight
F32
F32
[5120]
blk.28
blk.28.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.28.attn_k.weight
BF16
BF16
[5120, 1024]
blk.28.attn_k_norm.weight
F32
F32
[128]
blk.28.attn_norm.weight
F32
F32
[5120]
blk.28.attn_output.weight
BF16
BF16
[10240, 5120]
blk.28.attn_q.weight
BF16
BF16
[5120, 10240]
blk.28.attn_q_norm.weight
F32
F32
[128]
blk.28.attn_v.weight
BF16
BF16
[5120, 1024]
blk.28.attn_vres_lambda.weight
F32
F32
[1]
blk.28.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.28.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.28.ffn_norm.weight
F32
F32
[5120]
blk.28.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.28.post_attention_norm.weight
F32
F32
[5120]
blk.28.post_ffw_norm.weight
F32
F32
[5120]
blk.29
blk.29.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.29.attn_k.weight
BF16
BF16
[5120, 1024]
blk.29.attn_k_norm.weight
F32
F32
[128]
blk.29.attn_norm.weight
F32
F32
[5120]
blk.29.attn_output.weight
BF16
BF16
[10240, 5120]
blk.29.attn_q.weight
BF16
BF16
[5120, 10240]
blk.29.attn_q_norm.weight
F32
F32
[128]
blk.29.attn_v.weight
BF16
BF16
[5120, 1024]
blk.29.attn_vres_lambda.weight
F32
F32
[1]
blk.29.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.29.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.29.ffn_norm.weight
F32
F32
[5120]
blk.29.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.29.post_attention_norm.weight
F32
F32
[5120]
blk.29.post_ffw_norm.weight
F32
F32
[5120]
blk.30
blk.30.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.30.attn_k.weight
BF16
BF16
[5120, 1024]
blk.30.attn_k_norm.weight
F32
F32
[128]
blk.30.attn_norm.weight
F32
F32
[5120]
blk.30.attn_output.weight
BF16
BF16
[10240, 5120]
blk.30.attn_q.weight
BF16
BF16
[5120, 10240]
blk.30.attn_q_norm.weight
F32
F32
[128]
blk.30.attn_v.weight
BF16
BF16
[5120, 1024]
blk.30.attn_vres_lambda.weight
F32
F32
[1]
blk.30.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.30.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.30.ffn_norm.weight
F32
F32
[5120]
blk.30.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.30.post_attention_norm.weight
F32
F32
[5120]
blk.30.post_ffw_norm.weight
F32
F32
[5120]
blk.31
blk.31.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.31.attn_k.weight
BF16
BF16
[5120, 1024]
blk.31.attn_k_norm.weight
F32
F32
[128]
blk.31.attn_norm.weight
F32
F32
[5120]
blk.31.attn_output.weight
BF16
BF16
[10240, 5120]
blk.31.attn_q.weight
BF16
BF16
[5120, 10240]
blk.31.attn_q_norm.weight
F32
F32
[128]
blk.31.attn_v.weight
BF16
BF16
[5120, 1024]
blk.31.attn_vres_lambda.weight
F32
F32
[1]
blk.31.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.31.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.31.ffn_norm.weight
F32
F32
[5120]
blk.31.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.31.post_attention_norm.weight
F32
F32
[5120]
blk.31.post_ffw_norm.weight
F32
F32
[5120]
blk.32
blk.32.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.32.attn_k.weight
BF16
BF16
[5120, 1024]
blk.32.attn_k_norm.weight
F32
F32
[128]
blk.32.attn_norm.weight
F32
F32
[5120]
blk.32.attn_output.weight
BF16
BF16
[10240, 5120]
blk.32.attn_q.weight
BF16
BF16
[5120, 10240]
blk.32.attn_q_norm.weight
F32
F32
[128]
blk.32.attn_v.weight
BF16
BF16
[5120, 1024]
blk.32.attn_vres_lambda.weight
F32
F32
[1]
blk.32.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.32.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.32.ffn_norm.weight
F32
F32
[5120]
blk.32.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.32.post_attention_norm.weight
F32
F32
[5120]
blk.32.post_ffw_norm.weight
F32
F32
[5120]
blk.33
blk.33.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.33.attn_k.weight
BF16
BF16
[5120, 1024]
blk.33.attn_k_norm.weight
F32
F32
[128]
blk.33.attn_norm.weight
F32
F32
[5120]
blk.33.attn_output.weight
BF16
BF16
[10240, 5120]
blk.33.attn_q.weight
BF16
BF16
[5120, 10240]
blk.33.attn_q_norm.weight
F32
F32
[128]
blk.33.attn_v.weight
BF16
BF16
[5120, 1024]
blk.33.attn_vres_lambda.weight
F32
F32
[1]
blk.33.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.33.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.33.ffn_norm.weight
F32
F32
[5120]
blk.33.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.33.post_attention_norm.weight
F32
F32
[5120]
blk.33.post_ffw_norm.weight
F32
F32
[5120]
blk.34
blk.34.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.34.attn_k.weight
BF16
BF16
[5120, 1024]
blk.34.attn_k_norm.weight
F32
F32
[128]
blk.34.attn_norm.weight
F32
F32
[5120]
blk.34.attn_output.weight
BF16
BF16
[10240, 5120]
blk.34.attn_q.weight
BF16
BF16
[5120, 10240]
blk.34.attn_q_norm.weight
F32
F32
[128]
blk.34.attn_v.weight
BF16
BF16
[5120, 1024]
blk.34.attn_vres_lambda.weight
F32
F32
[1]
blk.34.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.34.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.34.ffn_norm.weight
F32
F32
[5120]
blk.34.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.34.post_attention_norm.weight
F32
F32
[5120]
blk.34.post_ffw_norm.weight
F32
F32
[5120]
blk.35
blk.35.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.35.attn_k.weight
BF16
BF16
[5120, 1024]
blk.35.attn_k_norm.weight
F32
F32
[128]
blk.35.attn_norm.weight
F32
F32
[5120]
blk.35.attn_output.weight
BF16
BF16
[10240, 5120]
blk.35.attn_q.weight
BF16
BF16
[5120, 10240]
blk.35.attn_q_norm.weight
F32
F32
[128]
blk.35.attn_v.weight
BF16
BF16
[5120, 1024]
blk.35.attn_vres_lambda.weight
F32
F32
[1]
blk.35.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.35.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.35.ffn_norm.weight
F32
F32
[5120]
blk.35.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.35.post_attention_norm.weight
F32
F32
[5120]
blk.35.post_ffw_norm.weight
F32
F32
[5120]
blk.36
blk.36.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.36.attn_k.weight
BF16
BF16
[5120, 1024]
blk.36.attn_k_norm.weight
F32
F32
[128]
blk.36.attn_norm.weight
F32
F32
[5120]
blk.36.attn_output.weight
BF16
BF16
[10240, 5120]
blk.36.attn_q.weight
BF16
BF16
[5120, 10240]
blk.36.attn_q_norm.weight
F32
F32
[128]
blk.36.attn_v.weight
BF16
BF16
[5120, 1024]
blk.36.attn_vres_lambda.weight
F32
F32
[1]
blk.36.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.36.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.36.ffn_norm.weight
F32
F32
[5120]
blk.36.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.36.post_attention_norm.weight
F32
F32
[5120]
blk.36.post_ffw_norm.weight
F32
F32
[5120]
blk.37
blk.37.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.37.attn_k.weight
BF16
BF16
[5120, 1024]
blk.37.attn_k_norm.weight
F32
F32
[128]
blk.37.attn_norm.weight
F32
F32
[5120]
blk.37.attn_output.weight
BF16
BF16
[10240, 5120]
blk.37.attn_q.weight
BF16
BF16
[5120, 10240]
blk.37.attn_q_norm.weight
F32
F32
[128]
blk.37.attn_v.weight
BF16
BF16
[5120, 1024]
blk.37.attn_vres_lambda.weight
F32
F32
[1]
blk.37.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.37.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.37.ffn_norm.weight
F32
F32
[5120]
blk.37.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.37.post_attention_norm.weight
F32
F32
[5120]
blk.37.post_ffw_norm.weight
F32
F32
[5120]
blk.38
blk.38.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.38.attn_k.weight
BF16
BF16
[5120, 1024]
blk.38.attn_k_norm.weight
F32
F32
[128]
blk.38.attn_norm.weight
F32
F32
[5120]
blk.38.attn_output.weight
BF16
BF16
[10240, 5120]
blk.38.attn_q.weight
BF16
BF16
[5120, 10240]
blk.38.attn_q_norm.weight
F32
F32
[128]
blk.38.attn_v.weight
BF16
BF16
[5120, 1024]
blk.38.attn_vres_lambda.weight
F32
F32
[1]
blk.38.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.38.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.38.ffn_norm.weight
F32
F32
[5120]
blk.38.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.38.post_attention_norm.weight
F32
F32
[5120]
blk.38.post_ffw_norm.weight
F32
F32
[5120]
blk.39
blk.39.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.39.attn_k.weight
BF16
BF16
[5120, 1024]
blk.39.attn_k_norm.weight
F32
F32
[128]
blk.39.attn_norm.weight
F32
F32
[5120]
blk.39.attn_output.weight
BF16
BF16
[10240, 5120]
blk.39.attn_q.weight
BF16
BF16
[5120, 10240]
blk.39.attn_q_norm.weight
F32
F32
[128]
blk.39.attn_v.weight
BF16
BF16
[5120, 1024]
blk.39.attn_vres_lambda.weight
F32
F32
[1]
blk.39.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.39.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.39.ffn_norm.weight
F32
F32
[5120]
blk.39.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.39.post_attention_norm.weight
F32
F32
[5120]
blk.39.post_ffw_norm.weight
F32
F32
[5120]
blk.40
blk.40.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.40.attn_k.weight
BF16
BF16
[5120, 1024]
blk.40.attn_k_norm.weight
F32
F32
[128]
blk.40.attn_norm.weight
F32
F32
[5120]
blk.40.attn_output.weight
BF16
BF16
[10240, 5120]
blk.40.attn_q.weight
BF16
BF16
[5120, 10240]
blk.40.attn_q_norm.weight
F32
F32
[128]
blk.40.attn_v.weight
BF16
BF16
[5120, 1024]
blk.40.attn_vres_lambda.weight
F32
F32
[1]
blk.40.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.40.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.40.ffn_norm.weight
F32
F32
[5120]
blk.40.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.40.post_attention_norm.weight
F32
F32
[5120]
blk.40.post_ffw_norm.weight
F32
F32
[5120]
blk.41
blk.41.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.41.attn_k.weight
BF16
BF16
[5120, 1024]
blk.41.attn_k_norm.weight
F32
F32
[128]
blk.41.attn_norm.weight
F32
F32
[5120]
blk.41.attn_output.weight
BF16
BF16
[10240, 5120]
blk.41.attn_q.weight
BF16
BF16
[5120, 10240]
blk.41.attn_q_norm.weight
F32
F32
[128]
blk.41.attn_v.weight
BF16
BF16
[5120, 1024]
blk.41.attn_vres_lambda.weight
F32
F32
[1]
blk.41.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.41.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.41.ffn_norm.weight
F32
F32
[5120]
blk.41.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.41.post_attention_norm.weight
F32
F32
[5120]
blk.41.post_ffw_norm.weight
F32
F32
[5120]
blk.42
blk.42.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.42.attn_k.weight
BF16
BF16
[5120, 1024]
blk.42.attn_k_norm.weight
F32
F32
[128]
blk.42.attn_norm.weight
F32
F32
[5120]
blk.42.attn_output.weight
BF16
BF16
[10240, 5120]
blk.42.attn_q.weight
BF16
BF16
[5120, 10240]
blk.42.attn_q_norm.weight
F32
F32
[128]
blk.42.attn_v.weight
BF16
BF16
[5120, 1024]
blk.42.attn_vres_lambda.weight
F32
F32
[1]
blk.42.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.42.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.42.ffn_norm.weight
F32
F32
[5120]
blk.42.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.42.post_attention_norm.weight
F32
F32
[5120]
blk.42.post_ffw_norm.weight
F32
F32
[5120]
blk.43
blk.43.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.43.attn_k.weight
BF16
BF16
[5120, 1024]
blk.43.attn_k_norm.weight
F32
F32
[128]
blk.43.attn_norm.weight
F32
F32
[5120]
blk.43.attn_output.weight
BF16
BF16
[10240, 5120]
blk.43.attn_q.weight
BF16
BF16
[5120, 10240]
blk.43.attn_q_norm.weight
F32
F32
[128]
blk.43.attn_v.weight
BF16
BF16
[5120, 1024]
blk.43.attn_vres_lambda.weight
F32
F32
[1]
blk.43.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.43.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.43.ffn_norm.weight
F32
F32
[5120]
blk.43.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.43.post_attention_norm.weight
F32
F32
[5120]
blk.43.post_ffw_norm.weight
F32
F32
[5120]
blk.44
blk.44.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.44.attn_k.weight
BF16
BF16
[5120, 1024]
blk.44.attn_k_norm.weight
F32
F32
[128]
blk.44.attn_norm.weight
F32
F32
[5120]
blk.44.attn_output.weight
BF16
BF16
[10240, 5120]
blk.44.attn_q.weight
BF16
BF16
[5120, 10240]
blk.44.attn_q_norm.weight
F32
F32
[128]
blk.44.attn_v.weight
BF16
BF16
[5120, 1024]
blk.44.attn_vres_lambda.weight
F32
F32
[1]
blk.44.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.44.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.44.ffn_norm.weight
F32
F32
[5120]
blk.44.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.44.post_attention_norm.weight
F32
F32
[5120]
blk.44.post_ffw_norm.weight
F32
F32
[5120]
blk.45
blk.45.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.45.attn_k.weight
BF16
BF16
[5120, 1024]
blk.45.attn_k_norm.weight
F32
F32
[128]
blk.45.attn_norm.weight
F32
F32
[5120]
blk.45.attn_output.weight
BF16
BF16
[10240, 5120]
blk.45.attn_q.weight
BF16
BF16
[5120, 10240]
blk.45.attn_q_norm.weight
F32
F32
[128]
blk.45.attn_v.weight
BF16
BF16
[5120, 1024]
blk.45.attn_vres_lambda.weight
F32
F32
[1]
blk.45.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.45.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.45.ffn_norm.weight
F32
F32
[5120]
blk.45.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.45.post_attention_norm.weight
F32
F32
[5120]
blk.45.post_ffw_norm.weight
F32
F32
[5120]
blk.46
blk.46.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.46.attn_k.weight
BF16
BF16
[5120, 1024]
blk.46.attn_k_norm.weight
F32
F32
[128]
blk.46.attn_norm.weight
F32
F32
[5120]
blk.46.attn_output.weight
BF16
BF16
[10240, 5120]
blk.46.attn_q.weight
BF16
BF16
[5120, 10240]
blk.46.attn_q_norm.weight
F32
F32
[128]
blk.46.attn_v.weight
BF16
BF16
[5120, 1024]
blk.46.attn_vres_lambda.weight
F32
F32
[1]
blk.46.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.46.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.46.ffn_norm.weight
F32
F32
[5120]
blk.46.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.46.post_attention_norm.weight
F32
F32
[5120]
blk.46.post_ffw_norm.weight
F32
F32
[5120]
blk.47
blk.47.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.47.attn_k.weight
BF16
BF16
[5120, 1024]
blk.47.attn_k_norm.weight
F32
F32
[128]
blk.47.attn_norm.weight
F32
F32
[5120]
blk.47.attn_output.weight
BF16
BF16
[10240, 5120]
blk.47.attn_q.weight
BF16
BF16
[5120, 10240]
blk.47.attn_q_norm.weight
F32
F32
[128]
blk.47.attn_v.weight
BF16
BF16
[5120, 1024]
blk.47.attn_vres_lambda.weight
F32
F32
[1]
blk.47.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.47.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.47.ffn_norm.weight
F32
F32
[5120]
blk.47.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.47.post_attention_norm.weight
F32
F32
[5120]
blk.47.post_ffw_norm.weight
F32
F32
[5120]
blk.48
blk.48.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.48.attn_k.weight
BF16
BF16
[5120, 1024]
blk.48.attn_k_norm.weight
F32
F32
[128]
blk.48.attn_norm.weight
F32
F32
[5120]
blk.48.attn_output.weight
BF16
BF16
[10240, 5120]
blk.48.attn_q.weight
BF16
BF16
[5120, 10240]
blk.48.attn_q_norm.weight
F32
F32
[128]
blk.48.attn_v.weight
BF16
BF16
[5120, 1024]
blk.48.attn_vres_lambda.weight
F32
F32
[1]
blk.48.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.48.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.48.ffn_norm.weight
F32
F32
[5120]
blk.48.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.48.post_attention_norm.weight
F32
F32
[5120]
blk.48.post_ffw_norm.weight
F32
F32
[5120]
blk.49
blk.49.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.49.attn_k.weight
BF16
BF16
[5120, 1024]
blk.49.attn_k_norm.weight
F32
F32
[128]
blk.49.attn_norm.weight
F32
F32
[5120]
blk.49.attn_output.weight
BF16
BF16
[10240, 5120]
blk.49.attn_q.weight
BF16
BF16
[5120, 10240]
blk.49.attn_q_norm.weight
F32
F32
[128]
blk.49.attn_v.weight
BF16
BF16
[5120, 1024]
blk.49.attn_vres_lambda.weight
F32
F32
[1]
blk.49.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.49.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.49.ffn_norm.weight
F32
F32
[5120]
blk.49.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.49.post_attention_norm.weight
F32
F32
[5120]
blk.49.post_ffw_norm.weight
F32
F32
[5120]
blk.50
blk.50.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.50.attn_k.weight
BF16
BF16
[5120, 1024]
blk.50.attn_k_norm.weight
F32
F32
[128]
blk.50.attn_norm.weight
F32
F32
[5120]
blk.50.attn_output.weight
BF16
BF16
[10240, 5120]
blk.50.attn_q.weight
BF16
BF16
[5120, 10240]
blk.50.attn_q_norm.weight
F32
F32
[128]
blk.50.attn_v.weight
BF16
BF16
[5120, 1024]
blk.50.attn_vres_lambda.weight
F32
F32
[1]
blk.50.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.50.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.50.ffn_norm.weight
F32
F32
[5120]
blk.50.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.50.post_attention_norm.weight
F32
F32
[5120]
blk.50.post_ffw_norm.weight
F32
F32
[5120]
blk.51
blk.51.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.51.attn_k.weight
BF16
BF16
[5120, 1024]
blk.51.attn_k_norm.weight
F32
F32
[128]
blk.51.attn_norm.weight
F32
F32
[5120]
blk.51.attn_output.weight
BF16
BF16
[10240, 5120]
blk.51.attn_q.weight
BF16
BF16
[5120, 10240]
blk.51.attn_q_norm.weight
F32
F32
[128]
blk.51.attn_v.weight
BF16
BF16
[5120, 1024]
blk.51.attn_vres_lambda.weight
F32
F32
[1]
blk.51.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.51.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.51.ffn_norm.weight
F32
F32
[5120]
blk.51.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.51.post_attention_norm.weight
F32
F32
[5120]
blk.51.post_ffw_norm.weight
F32
F32
[5120]
blk.52
blk.52.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.52.attn_k.weight
BF16
BF16
[5120, 1024]
blk.52.attn_k_norm.weight
F32
F32
[128]
blk.52.attn_norm.weight
F32
F32
[5120]
blk.52.attn_output.weight
BF16
BF16
[10240, 5120]
blk.52.attn_q.weight
BF16
BF16
[5120, 10240]
blk.52.attn_q_norm.weight
F32
F32
[128]
blk.52.attn_v.weight
BF16
BF16
[5120, 1024]
blk.52.attn_vres_lambda.weight
F32
F32
[1]
blk.52.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.52.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.52.ffn_norm.weight
F32
F32
[5120]
blk.52.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.52.post_attention_norm.weight
F32
F32
[5120]
blk.52.post_ffw_norm.weight
F32
F32
[5120]
blk.53
blk.53.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.53.attn_k.weight
BF16
BF16
[5120, 1024]
blk.53.attn_k_norm.weight
F32
F32
[128]
blk.53.attn_norm.weight
F32
F32
[5120]
blk.53.attn_output.weight
BF16
BF16
[10240, 5120]
blk.53.attn_q.weight
BF16
BF16
[5120, 10240]
blk.53.attn_q_norm.weight
F32
F32
[128]
blk.53.attn_v.weight
BF16
BF16
[5120, 1024]
blk.53.attn_vres_lambda.weight
F32
F32
[1]
blk.53.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.53.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.53.ffn_norm.weight
F32
F32
[5120]
blk.53.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.53.post_attention_norm.weight
F32
F32
[5120]
blk.53.post_ffw_norm.weight
F32
F32
[5120]
blk.54
blk.54.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.54.attn_k.weight
BF16
BF16
[5120, 1024]
blk.54.attn_k_norm.weight
F32
F32
[128]
blk.54.attn_norm.weight
F32
F32
[5120]
blk.54.attn_output.weight
BF16
BF16
[10240, 5120]
blk.54.attn_q.weight
BF16
BF16
[5120, 10240]
blk.54.attn_q_norm.weight
F32
F32
[128]
blk.54.attn_v.weight
BF16
BF16
[5120, 1024]
blk.54.attn_vres_lambda.weight
F32
F32
[1]
blk.54.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.54.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.54.ffn_norm.weight
F32
F32
[5120]
blk.54.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.54.post_attention_norm.weight
F32
F32
[5120]
blk.54.post_ffw_norm.weight
F32
F32
[5120]
blk.55
blk.55.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.55.attn_k.weight
BF16
BF16
[5120, 1024]
blk.55.attn_k_norm.weight
F32
F32
[128]
blk.55.attn_norm.weight
F32
F32
[5120]
blk.55.attn_output.weight
BF16
BF16
[10240, 5120]
blk.55.attn_q.weight
BF16
BF16
[5120, 10240]
blk.55.attn_q_norm.weight
F32
F32
[128]
blk.55.attn_v.weight
BF16
BF16
[5120, 1024]
blk.55.attn_vres_lambda.weight
F32
F32
[1]
blk.55.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.55.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.55.ffn_norm.weight
F32
F32
[5120]
blk.55.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.55.post_attention_norm.weight
F32
F32
[5120]
blk.55.post_ffw_norm.weight
F32
F32
[5120]
blk.56
blk.56.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.56.attn_k.weight
BF16
BF16
[5120, 1024]
blk.56.attn_k_norm.weight
F32
F32
[128]
blk.56.attn_norm.weight
F32
F32
[5120]
blk.56.attn_output.weight
BF16
BF16
[10240, 5120]
blk.56.attn_q.weight
BF16
BF16
[5120, 10240]
blk.56.attn_q_norm.weight
F32
F32
[128]
blk.56.attn_v.weight
BF16
BF16
[5120, 1024]
blk.56.attn_vres_lambda.weight
F32
F32
[1]
blk.56.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.56.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.56.ffn_norm.weight
F32
F32
[5120]
blk.56.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.56.post_attention_norm.weight
F32
F32
[5120]
blk.56.post_ffw_norm.weight
F32
F32
[5120]
blk.57
blk.57.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.57.attn_k.weight
BF16
BF16
[5120, 1024]
blk.57.attn_k_norm.weight
F32
F32
[128]
blk.57.attn_norm.weight
F32
F32
[5120]
blk.57.attn_output.weight
BF16
BF16
[10240, 5120]
blk.57.attn_q.weight
BF16
BF16
[5120, 10240]
blk.57.attn_q_norm.weight
F32
F32
[128]
blk.57.attn_v.weight
BF16
BF16
[5120, 1024]
blk.57.attn_vres_lambda.weight
F32
F32
[1]
blk.57.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.57.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.57.ffn_norm.weight
F32
F32
[5120]
blk.57.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.57.post_attention_norm.weight
F32
F32
[5120]
blk.57.post_ffw_norm.weight
F32
F32
[5120]
blk.58
blk.58.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.58.attn_k.weight
BF16
BF16
[5120, 1024]
blk.58.attn_k_norm.weight
F32
F32
[128]
blk.58.attn_norm.weight
F32
F32
[5120]
blk.58.attn_output.weight
BF16
BF16
[10240, 5120]
blk.58.attn_q.weight
BF16
BF16
[5120, 10240]
blk.58.attn_q_norm.weight
F32
F32
[128]
blk.58.attn_v.weight
BF16
BF16
[5120, 1024]
blk.58.attn_vres_lambda.weight
F32
F32
[1]
blk.58.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.58.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.58.ffn_norm.weight
F32
F32
[5120]
blk.58.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.58.post_attention_norm.weight
F32
F32
[5120]
blk.58.post_ffw_norm.weight
F32
F32
[5120]
blk.59
blk.59.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.59.attn_k.weight
BF16
BF16
[5120, 1024]
blk.59.attn_k_norm.weight
F32
F32
[128]
blk.59.attn_norm.weight
F32
F32
[5120]
blk.59.attn_output.weight
BF16
BF16
[10240, 5120]
blk.59.attn_q.weight
BF16
BF16
[5120, 10240]
blk.59.attn_q_norm.weight
F32
F32
[128]
blk.59.attn_v.weight
BF16
BF16
[5120, 1024]
blk.59.attn_vres_lambda.weight
F32
F32
[1]
blk.59.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.59.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.59.ffn_norm.weight
F32
F32
[5120]
blk.59.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.59.post_attention_norm.weight
F32
F32
[5120]
blk.59.post_ffw_norm.weight
F32
F32
[5120]
blk.60
blk.60.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.60.attn_k.weight
BF16
BF16
[5120, 1024]
blk.60.attn_k_norm.weight
F32
F32
[128]
blk.60.attn_norm.weight
F32
F32
[5120]
blk.60.attn_output.weight
BF16
BF16
[10240, 5120]
blk.60.attn_q.weight
BF16
BF16
[5120, 10240]
blk.60.attn_q_norm.weight
F32
F32
[128]
blk.60.attn_v.weight
BF16
BF16
[5120, 1024]
blk.60.attn_vres_lambda.weight
F32
F32
[1]
blk.60.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.60.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.60.ffn_norm.weight
F32
F32
[5120]
blk.60.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.60.post_attention_norm.weight
F32
F32
[5120]
blk.60.post_ffw_norm.weight
F32
F32
[5120]
blk.61
blk.61.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.61.attn_k.weight
BF16
BF16
[5120, 1024]
blk.61.attn_k_norm.weight
F32
F32
[128]
blk.61.attn_norm.weight
F32
F32
[5120]
blk.61.attn_output.weight
BF16
BF16
[10240, 5120]
blk.61.attn_q.weight
BF16
BF16
[5120, 10240]
blk.61.attn_q_norm.weight
F32
F32
[128]
blk.61.attn_v.weight
BF16
BF16
[5120, 1024]
blk.61.attn_vres_lambda.weight
F32
F32
[1]
blk.61.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.61.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.61.ffn_norm.weight
F32
F32
[5120]
blk.61.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.61.post_attention_norm.weight
F32
F32
[5120]
blk.61.post_ffw_norm.weight
F32
F32
[5120]
blk.62
blk.62.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.62.attn_k.weight
BF16
BF16
[5120, 1024]
blk.62.attn_k_norm.weight
F32
F32
[128]
blk.62.attn_norm.weight
F32
F32
[5120]
blk.62.attn_output.weight
BF16
BF16
[10240, 5120]
blk.62.attn_q.weight
BF16
BF16
[5120, 10240]
blk.62.attn_q_norm.weight
F32
F32
[128]
blk.62.attn_v.weight
BF16
BF16
[5120, 1024]
blk.62.attn_vres_lambda.weight
F32
F32
[1]
blk.62.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.62.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.62.ffn_norm.weight
F32
F32
[5120]
blk.62.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.62.post_attention_norm.weight
F32
F32
[5120]
blk.62.post_ffw_norm.weight
F32
F32
[5120]
blk.63
blk.63.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.63.attn_k.weight
BF16
BF16
[5120, 1024]
blk.63.attn_k_norm.weight
F32
F32
[128]
blk.63.attn_norm.weight
F32
F32
[5120]
blk.63.attn_output.weight
BF16
BF16
[10240, 5120]
blk.63.attn_q.weight
BF16
BF16
[5120, 10240]
blk.63.attn_q_norm.weight
F32
F32
[128]
blk.63.attn_v.weight
BF16
BF16
[5120, 1024]
blk.63.attn_vres_lambda.weight
F32
F32
[1]
blk.63.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.63.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.63.ffn_norm.weight
F32
F32
[5120]
blk.63.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.63.post_attention_norm.weight
F32
F32
[5120]
blk.63.post_ffw_norm.weight
F32
F32
[5120]
blk.64
blk.64.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.64.attn_k.weight
BF16
BF16
[5120, 1024]
blk.64.attn_k_norm.weight
F32
F32
[128]
blk.64.attn_norm.weight
F32
F32
[5120]
blk.64.attn_output.weight
BF16
BF16
[10240, 5120]
blk.64.attn_q.weight
BF16
BF16
[5120, 10240]
blk.64.attn_q_norm.weight
F32
F32
[128]
blk.64.attn_v.weight
BF16
BF16
[5120, 1024]
blk.64.attn_vres_lambda.weight
F32
F32
[1]
blk.64.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.64.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.64.ffn_norm.weight
F32
F32
[5120]
blk.64.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.64.post_attention_norm.weight
F32
F32
[5120]
blk.64.post_ffw_norm.weight
F32
F32
[5120]
blk.65
blk.65.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.65.attn_k.weight
BF16
BF16
[5120, 1024]
blk.65.attn_k_norm.weight
F32
F32
[128]
blk.65.attn_norm.weight
F32
F32
[5120]
blk.65.attn_output.weight
BF16
BF16
[10240, 5120]
blk.65.attn_q.weight
BF16
BF16
[5120, 10240]
blk.65.attn_q_norm.weight
F32
F32
[128]
blk.65.attn_v.weight
BF16
BF16
[5120, 1024]
blk.65.attn_vres_lambda.weight
F32
F32
[1]
blk.65.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.65.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.65.ffn_norm.weight
F32
F32
[5120]
blk.65.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.65.post_attention_norm.weight
F32
F32
[5120]
blk.65.post_ffw_norm.weight
F32
F32
[5120]
blk.66
blk.66.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.66.attn_k.weight
BF16
BF16
[5120, 1024]
blk.66.attn_k_norm.weight
F32
F32
[128]
blk.66.attn_norm.weight
F32
F32
[5120]
blk.66.attn_output.weight
BF16
BF16
[10240, 5120]
blk.66.attn_q.weight
BF16
BF16
[5120, 10240]
blk.66.attn_q_norm.weight
F32
F32
[128]
blk.66.attn_v.weight
BF16
BF16
[5120, 1024]
blk.66.attn_vres_lambda.weight
F32
F32
[1]
blk.66.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.66.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.66.ffn_norm.weight
F32
F32
[5120]
blk.66.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.66.post_attention_norm.weight
F32
F32
[5120]
blk.66.post_ffw_norm.weight
F32
F32
[5120]
blk.67
blk.67.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.67.attn_k.weight
BF16
BF16
[5120, 1024]
blk.67.attn_k_norm.weight
F32
F32
[128]
blk.67.attn_norm.weight
F32
F32
[5120]
blk.67.attn_output.weight
BF16
BF16
[10240, 5120]
blk.67.attn_q.weight
BF16
BF16
[5120, 10240]
blk.67.attn_q_norm.weight
F32
F32
[128]
blk.67.attn_v.weight
BF16
BF16
[5120, 1024]
blk.67.attn_vres_lambda.weight
F32
F32
[1]
blk.67.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.67.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.67.ffn_norm.weight
F32
F32
[5120]
blk.67.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.67.post_attention_norm.weight
F32
F32
[5120]
blk.67.post_ffw_norm.weight
F32
F32
[5120]
blk.68
blk.68.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.68.attn_k.weight
BF16
BF16
[5120, 1024]
blk.68.attn_k_norm.weight
F32
F32
[128]
blk.68.attn_norm.weight
F32
F32
[5120]
blk.68.attn_output.weight
BF16
BF16
[10240, 5120]
blk.68.attn_q.weight
BF16
BF16
[5120, 10240]
blk.68.attn_q_norm.weight
F32
F32
[128]
blk.68.attn_v.weight
BF16
BF16
[5120, 1024]
blk.68.attn_vres_lambda.weight
F32
F32
[1]
blk.68.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.68.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.68.ffn_norm.weight
F32
F32
[5120]
blk.68.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.68.post_attention_norm.weight
F32
F32
[5120]
blk.68.post_ffw_norm.weight
F32
F32
[5120]
blk.69
blk.69.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.69.attn_k.weight
BF16
BF16
[5120, 1024]
blk.69.attn_k_norm.weight
F32
F32
[128]
blk.69.attn_norm.weight
F32
F32
[5120]
blk.69.attn_output.weight
BF16
BF16
[10240, 5120]
blk.69.attn_q.weight
BF16
BF16
[5120, 10240]
blk.69.attn_q_norm.weight
F32
F32
[128]
blk.69.attn_v.weight
BF16
BF16
[5120, 1024]
blk.69.attn_vres_lambda.weight
F32
F32
[1]
blk.69.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.69.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.69.ffn_norm.weight
F32
F32
[5120]
blk.69.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.69.post_attention_norm.weight
F32
F32
[5120]
blk.69.post_ffw_norm.weight
F32
F32
[5120]
blk.70
blk.70.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.70.attn_k.weight
BF16
BF16
[5120, 1024]
blk.70.attn_k_norm.weight
F32
F32
[128]
blk.70.attn_norm.weight
F32
F32
[5120]
blk.70.attn_output.weight
BF16
BF16
[10240, 5120]
blk.70.attn_q.weight
BF16
BF16
[5120, 10240]
blk.70.attn_q_norm.weight
F32
F32
[128]
blk.70.attn_v.weight
BF16
BF16
[5120, 1024]
blk.70.attn_vres_lambda.weight
F32
F32
[1]
blk.70.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.70.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.70.ffn_norm.weight
F32
F32
[5120]
blk.70.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.70.post_attention_norm.weight
F32
F32
[5120]
blk.70.post_ffw_norm.weight
F32
F32
[5120]
blk.71
blk.71.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.71.attn_k.weight
BF16
BF16
[5120, 1024]
blk.71.attn_k_norm.weight
F32
F32
[128]
blk.71.attn_norm.weight
F32
F32
[5120]
blk.71.attn_output.weight
BF16
BF16
[10240, 5120]
blk.71.attn_q.weight
BF16
BF16
[5120, 10240]
blk.71.attn_q_norm.weight
F32
F32
[128]
blk.71.attn_v.weight
BF16
BF16
[5120, 1024]
blk.71.attn_vres_lambda.weight
F32
F32
[1]
blk.71.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.71.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.71.ffn_norm.weight
F32
F32
[5120]
blk.71.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.71.post_attention_norm.weight
F32
F32
[5120]
blk.71.post_ffw_norm.weight
F32
F32
[5120]
blk.72
blk.72.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.72.attn_k.weight
BF16
BF16
[5120, 1024]
blk.72.attn_k_norm.weight
F32
F32
[128]
blk.72.attn_norm.weight
F32
F32
[5120]
blk.72.attn_output.weight
BF16
BF16
[10240, 5120]
blk.72.attn_q.weight
BF16
BF16
[5120, 10240]
blk.72.attn_q_norm.weight
F32
F32
[128]
blk.72.attn_v.weight
BF16
BF16
[5120, 1024]
blk.72.attn_vres_lambda.weight
F32
F32
[1]
blk.72.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.72.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.72.ffn_norm.weight
F32
F32
[5120]
blk.72.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.72.post_attention_norm.weight
F32
F32
[5120]
blk.72.post_ffw_norm.weight
F32
F32
[5120]
blk.73
blk.73.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.73.attn_k.weight
BF16
BF16
[5120, 1024]
blk.73.attn_k_norm.weight
F32
F32
[128]
blk.73.attn_norm.weight
F32
F32
[5120]
blk.73.attn_output.weight
BF16
BF16
[10240, 5120]
blk.73.attn_q.weight
BF16
BF16
[5120, 10240]
blk.73.attn_q_norm.weight
F32
F32
[128]
blk.73.attn_v.weight
BF16
BF16
[5120, 1024]
blk.73.attn_vres_lambda.weight
F32
F32
[1]
blk.73.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.73.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.73.ffn_norm.weight
F32
F32
[5120]
blk.73.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.73.post_attention_norm.weight
F32
F32
[5120]
blk.73.post_ffw_norm.weight
F32
F32
[5120]
blk.74
blk.74.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.74.attn_k.weight
BF16
BF16
[5120, 1024]
blk.74.attn_k_norm.weight
F32
F32
[128]
blk.74.attn_norm.weight
F32
F32
[5120]
blk.74.attn_output.weight
BF16
BF16
[10240, 5120]
blk.74.attn_q.weight
BF16
BF16
[5120, 10240]
blk.74.attn_q_norm.weight
F32
F32
[128]
blk.74.attn_v.weight
BF16
BF16
[5120, 1024]
blk.74.attn_vres_lambda.weight
F32
F32
[1]
blk.74.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.74.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.74.ffn_norm.weight
F32
F32
[5120]
blk.74.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.74.post_attention_norm.weight
F32
F32
[5120]
blk.74.post_ffw_norm.weight
F32
F32
[5120]
blk.75
blk.75.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.75.attn_k.weight
BF16
BF16
[5120, 1024]
blk.75.attn_k_norm.weight
F32
F32
[128]
blk.75.attn_norm.weight
F32
F32
[5120]
blk.75.attn_output.weight
BF16
BF16
[10240, 5120]
blk.75.attn_q.weight
BF16
BF16
[5120, 10240]
blk.75.attn_q_norm.weight
F32
F32
[128]
blk.75.attn_v.weight
BF16
BF16
[5120, 1024]
blk.75.attn_vres_lambda.weight
F32
F32
[1]
blk.75.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.75.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.75.ffn_norm.weight
F32
F32
[5120]
blk.75.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.75.post_attention_norm.weight
F32
F32
[5120]
blk.75.post_ffw_norm.weight
F32
F32
[5120]
blk.76
blk.76.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.76.attn_k.weight
BF16
BF16
[5120, 1024]
blk.76.attn_k_norm.weight
F32
F32
[128]
blk.76.attn_norm.weight
F32
F32
[5120]
blk.76.attn_output.weight
BF16
BF16
[10240, 5120]
blk.76.attn_q.weight
BF16
BF16
[5120, 10240]
blk.76.attn_q_norm.weight
F32
F32
[128]
blk.76.attn_v.weight
BF16
BF16
[5120, 1024]
blk.76.attn_vres_lambda.weight
F32
F32
[1]
blk.76.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.76.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.76.ffn_norm.weight
F32
F32
[5120]
blk.76.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.76.post_attention_norm.weight
F32
F32
[5120]
blk.76.post_ffw_norm.weight
F32
F32
[5120]
blk.77
blk.77.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.77.attn_k.weight
BF16
BF16
[5120, 1024]
blk.77.attn_k_norm.weight
F32
F32
[128]
blk.77.attn_norm.weight
F32
F32
[5120]
blk.77.attn_output.weight
BF16
BF16
[10240, 5120]
blk.77.attn_q.weight
BF16
BF16
[5120, 10240]
blk.77.attn_q_norm.weight
F32
F32
[128]
blk.77.attn_v.weight
BF16
BF16
[5120, 1024]
blk.77.attn_vres_lambda.weight
F32
F32
[1]
blk.77.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.77.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.77.ffn_norm.weight
F32
F32
[5120]
blk.77.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.77.post_attention_norm.weight
F32
F32
[5120]
blk.77.post_ffw_norm.weight
F32
F32
[5120]
blk.78
blk.78.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.78.attn_k.weight
BF16
BF16
[5120, 1024]
blk.78.attn_k_norm.weight
F32
F32
[128]
blk.78.attn_norm.weight
F32
F32
[5120]
blk.78.attn_output.weight
BF16
BF16
[10240, 5120]
blk.78.attn_q.weight
BF16
BF16
[5120, 10240]
blk.78.attn_q_norm.weight
F32
F32
[128]
blk.78.attn_v.weight
BF16
BF16
[5120, 1024]
blk.78.attn_vres_lambda.weight
F32
F32
[1]
blk.78.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.78.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.78.ffn_norm.weight
F32
F32
[5120]
blk.78.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.78.post_attention_norm.weight
F32
F32
[5120]
blk.78.post_ffw_norm.weight
F32
F32
[5120]
blk.79
blk.79.attn_gate.weight
BF16
BF16
[5120, 10240]
blk.79.attn_k.weight
BF16
BF16
[5120, 1024]
blk.79.attn_k_norm.weight
F32
F32
[128]
blk.79.attn_norm.weight
F32
F32
[5120]
blk.79.attn_output.weight
BF16
BF16
[10240, 5120]
blk.79.attn_q.weight
BF16
BF16
[5120, 10240]
blk.79.attn_q_norm.weight
F32
F32
[128]
blk.79.attn_v.weight
BF16
BF16
[5120, 1024]
blk.79.attn_vres_lambda.weight
F32
F32
[1]
blk.79.ffn_down.weight
BF16
BF16
[40960, 5120]
blk.79.ffn_gate.weight
BF16
BF16
[5120, 40960]
blk.79.ffn_norm.weight
F32
F32
[5120]
blk.79.ffn_up.weight
BF16
BF16
[5120, 40960]
blk.79.post_attention_norm.weight
F32
F32
[5120]
blk.79.post_ffw_norm.weight
F32
F32
[5120]
output.weight
BF16
BF16
[5120, 100352]
output_norm.weight
F32
F32
[5120]