GitHub
186.6K
Download
Models
Docs
Pricing
Sign in
Download
Models
Docs
Pricing
Sign in
siddhu19
/
flint-unsloth
:q2_xl
50
Downloads
Updated
2 weeks ago
GGUF quantizations of Flint 39.8B by Recursive AI for coding agents, with Unsloth's Dynamic 3.0 quantization method
GGUF quantizations of Flint 39.8B by Recursive AI for coding agents, with Unsloth's Dynamic 3.0 quantization method
Cancel
flint-unsloth:q2_xl
...
/
model
7f43dd99e7d5 · 14GB
Metadata
general.architecture
flint
flint
general.file_type
Q2_K
Q2_K
flint.attention.head_count
32
32
flint.attention.head_count_kv
4
4
flint.attention.key_length
128
128
flint.attention.layer_norm_rms_epsilon
1e-06
1e-06
flint.attention.rope_pattern
[false, false, true, true, true, ...]
[false, false, true, true, true, ...]
flint.attention.sliding_window
513
513
flint.attention.sliding_window_pattern
[false, false, true, true, true, ...]
[false, false, true, true, true, ...]
flint.attention.value_length
128
128
flint.block_count
49
49
flint.context_length
262144
262144
flint.embedding_length
2048
2048
flint.expert_count
256
256
flint.expert_feed_forward_length
512
512
flint.expert_gating_func
2
2
flint.expert_shared_count
2
2
flint.expert_shared_feed_forward_length
1024
1024
flint.expert_used_count
6
6
flint.expert_weights_norm
true
true
flint.expert_weights_scale
1
1
flint.feed_forward_length
6144
6144
flint.leading_dense_block_count
2
2
flint.rope.freq_base
10000
10000
tokenizer.ggml.add_bos_token
false
false
tokenizer.ggml.add_eos_token
false
false
tokenizer.ggml.bos_token_id
150000
150000
tokenizer.ggml.eog_token_ids
[150001, 150003, 150005, 150018]
[150001, 150003, 150005, 150018]
tokenizer.ggml.eos_token_id
150001
150001
tokenizer.ggml.merges
[Ġ Ġ, Ġ t, ĠĠ ĠĠ, i n, Ġ a, ...]
[Ġ Ġ, Ġ t, ĠĠ ĠĠ, i n, Ġ a, ...]
tokenizer.ggml.model
gpt2
gpt2
tokenizer.ggml.padding_token_id
150001
150001
tokenizer.ggml.pre
flint
flint
tokenizer.ggml.token_type
[1, 1, 1, 1, 1, ...]
[1, 1, 1, 1, 1, ...]
tokenizer.ggml.tokens
[!, ", #, $, %, ...]
[!, ", #, $, %, ...]
quantize.imatrix.chunks_count
687
687
quantize.imatrix.dataset
unsloth_calibration_flint-preview.txt
unsloth_calibration_flint-preview.txt
quantize.imatrix.entries_count
531
531
quantize.imatrix.file
flint-preview-GGUF/imatrix_unsloth.gguf_file
flint-preview-GGUF/imatrix_unsloth.gguf_file
Tensor
Name
Type
Shape
token_embd.weight
Q4_K
Q4_K
[2048, 150144]
blk.0
blk.0.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.0.attn_k_norm.weight
F32
F32
[128]
blk.0.attn_norm.weight
F32
F32
[2048]
blk.0.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.0.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.0.attn_q_norm.weight
F32
F32
[128]
blk.0.attn_sinks.weight
F32
F32
[32]
blk.0.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.0.ffn_down.weight
Q5_K
Q5_K
[6144, 2048]
blk.0.ffn_gate.weight
Q5_K
Q5_K
[2048, 6144]
blk.0.ffn_norm.weight
F32
F32
[2048]
blk.0.ffn_up.weight
Q5_K
Q5_K
[2048, 6144]
blk.1
blk.1.attn_k.weight
Q6_K
Q6_K
[2048, 512]
blk.1.attn_k_norm.weight
F32
F32
[128]
blk.1.attn_norm.weight
F32
F32
[2048]
blk.1.attn_output.weight
Q6_K
Q6_K
[4096, 2048]
blk.1.attn_q.weight
Q6_K
Q6_K
[2048, 4096]
blk.1.attn_q_norm.weight
F32
F32
[128]
blk.1.attn_sinks.weight
F32
F32
[32]
blk.1.attn_v.weight
Q6_K
Q6_K
[2048, 512]
blk.1.ffn_down.weight
Q5_K
Q5_K
[6144, 2048]
blk.1.ffn_gate.weight
Q5_K
Q5_K
[2048, 6144]
blk.1.ffn_norm.weight
F32
F32
[2048]
blk.1.ffn_up.weight
Q5_K
Q5_K
[2048, 6144]
blk.2
blk.2.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.2.attn_k_norm.weight
F32
F32
[128]
blk.2.attn_norm.weight
F32
F32
[2048]
blk.2.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.2.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.2.attn_q_norm.weight
F32
F32
[128]
blk.2.attn_sinks.weight
F32
F32
[32]
blk.2.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.2.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.2.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.2.ffn_gate_exps.weight
IQ1_S
IQ1_S
[2048, 512, 256]
blk.2.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.2.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.2.ffn_norm.weight
F32
F32
[2048]
blk.2.ffn_up_exps.weight
IQ1_S
IQ1_S
[2048, 512, 256]
blk.2.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.3
blk.3.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.3.attn_k_norm.weight
F32
F32
[128]
blk.3.attn_norm.weight
F32
F32
[2048]
blk.3.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.3.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.3.attn_q_norm.weight
F32
F32
[128]
blk.3.attn_sinks.weight
F32
F32
[32]
blk.3.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.3.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.3.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.3.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.3.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.3.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.3.ffn_norm.weight
F32
F32
[2048]
blk.3.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.3.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.4
blk.4.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.4.attn_k_norm.weight
F32
F32
[128]
blk.4.attn_norm.weight
F32
F32
[2048]
blk.4.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.4.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.4.attn_q_norm.weight
F32
F32
[128]
blk.4.attn_sinks.weight
F32
F32
[32]
blk.4.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.4.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.4.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.4.ffn_gate_exps.weight
IQ2_XS
IQ2_XS
[2048, 512, 256]
blk.4.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.4.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.4.ffn_norm.weight
F32
F32
[2048]
blk.4.ffn_up_exps.weight
IQ2_XS
IQ2_XS
[2048, 512, 256]
blk.4.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.5
blk.5.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.5.attn_k_norm.weight
F32
F32
[128]
blk.5.attn_norm.weight
F32
F32
[2048]
blk.5.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.5.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.5.attn_q_norm.weight
F32
F32
[128]
blk.5.attn_sinks.weight
F32
F32
[32]
blk.5.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.5.ffn_down_exps.weight
IQ3_S
IQ3_S
[512, 2048, 256]
blk.5.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.5.ffn_gate_exps.weight
IQ2_XS
IQ2_XS
[2048, 512, 256]
blk.5.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.5.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.5.ffn_norm.weight
F32
F32
[2048]
blk.5.ffn_up_exps.weight
IQ2_XS
IQ2_XS
[2048, 512, 256]
blk.5.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.6
blk.6.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.6.attn_k_norm.weight
F32
F32
[128]
blk.6.attn_norm.weight
F32
F32
[2048]
blk.6.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.6.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.6.attn_q_norm.weight
F32
F32
[128]
blk.6.attn_sinks.weight
F32
F32
[32]
blk.6.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.6.ffn_down_exps.weight
IQ3_S
IQ3_S
[512, 2048, 256]
blk.6.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.6.ffn_gate_exps.weight
IQ2_XS
IQ2_XS
[2048, 512, 256]
blk.6.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.6.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.6.ffn_norm.weight
F32
F32
[2048]
blk.6.ffn_up_exps.weight
IQ2_XS
IQ2_XS
[2048, 512, 256]
blk.6.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.7
blk.7.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.7.attn_k_norm.weight
F32
F32
[128]
blk.7.attn_norm.weight
F32
F32
[2048]
blk.7.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.7.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.7.attn_q_norm.weight
F32
F32
[128]
blk.7.attn_sinks.weight
F32
F32
[32]
blk.7.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.7.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[512, 2048, 256]
blk.7.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.7.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.7.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.7.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.7.ffn_norm.weight
F32
F32
[2048]
blk.7.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.7.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.8
blk.8.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.8.attn_k_norm.weight
F32
F32
[128]
blk.8.attn_norm.weight
F32
F32
[2048]
blk.8.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.8.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.8.attn_q_norm.weight
F32
F32
[128]
blk.8.attn_sinks.weight
F32
F32
[32]
blk.8.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.8.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[512, 2048, 256]
blk.8.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.8.ffn_gate_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.8.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.8.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.8.ffn_norm.weight
F32
F32
[2048]
blk.8.ffn_up_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.8.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.9
blk.9.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.9.attn_k_norm.weight
F32
F32
[128]
blk.9.attn_norm.weight
F32
F32
[2048]
blk.9.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.9.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.9.attn_q_norm.weight
F32
F32
[128]
blk.9.attn_sinks.weight
F32
F32
[32]
blk.9.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.9.ffn_down_exps.weight
IQ3_S
IQ3_S
[512, 2048, 256]
blk.9.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.9.ffn_gate_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.9.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.9.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.9.ffn_norm.weight
F32
F32
[2048]
blk.9.ffn_up_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.9.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.10
blk.10.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.10.attn_k_norm.weight
F32
F32
[128]
blk.10.attn_norm.weight
F32
F32
[2048]
blk.10.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.10.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.10.attn_q_norm.weight
F32
F32
[128]
blk.10.attn_sinks.weight
F32
F32
[32]
blk.10.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.10.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[512, 2048, 256]
blk.10.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.10.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.10.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.10.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.10.ffn_norm.weight
F32
F32
[2048]
blk.10.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.10.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.11
blk.11.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.11.attn_k_norm.weight
F32
F32
[128]
blk.11.attn_norm.weight
F32
F32
[2048]
blk.11.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.11.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.11.attn_q_norm.weight
F32
F32
[128]
blk.11.attn_sinks.weight
F32
F32
[32]
blk.11.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.11.ffn_down_exps.weight
IQ4_XS
IQ4_XS
[512, 2048, 256]
blk.11.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.11.ffn_gate_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.11.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.11.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.11.ffn_norm.weight
F32
F32
[2048]
blk.11.ffn_up_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.11.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.12
blk.12.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.12.attn_k_norm.weight
F32
F32
[128]
blk.12.attn_norm.weight
F32
F32
[2048]
blk.12.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.12.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.12.attn_q_norm.weight
F32
F32
[128]
blk.12.attn_sinks.weight
F32
F32
[32]
blk.12.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.12.ffn_down_exps.weight
IQ3_S
IQ3_S
[512, 2048, 256]
blk.12.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.12.ffn_gate_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.12.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.12.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.12.ffn_norm.weight
F32
F32
[2048]
blk.12.ffn_up_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.12.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.13
blk.13.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.13.attn_k_norm.weight
F32
F32
[128]
blk.13.attn_norm.weight
F32
F32
[2048]
blk.13.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.13.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.13.attn_q_norm.weight
F32
F32
[128]
blk.13.attn_sinks.weight
F32
F32
[32]
blk.13.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.13.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.13.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.13.ffn_gate_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.13.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.13.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.13.ffn_norm.weight
F32
F32
[2048]
blk.13.ffn_up_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.13.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.14
blk.14.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.14.attn_k_norm.weight
F32
F32
[128]
blk.14.attn_norm.weight
F32
F32
[2048]
blk.14.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.14.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.14.attn_q_norm.weight
F32
F32
[128]
blk.14.attn_sinks.weight
F32
F32
[32]
blk.14.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.14.ffn_down_exps.weight
IQ3_S
IQ3_S
[512, 2048, 256]
blk.14.ffn_down_shexp.weight
Q6_K
Q6_K
[1024, 2048]
blk.14.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.14.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.14.ffn_gate_shexp.weight
Q6_K
Q6_K
[2048, 1024]
blk.14.ffn_norm.weight
F32
F32
[2048]
blk.14.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.14.ffn_up_shexp.weight
Q6_K
Q6_K
[2048, 1024]
blk.15
blk.15.attn_k.weight
Q6_K
Q6_K
[2048, 512]
blk.15.attn_k_norm.weight
F32
F32
[128]
blk.15.attn_norm.weight
F32
F32
[2048]
blk.15.attn_output.weight
Q6_K
Q6_K
[4096, 2048]
blk.15.attn_q.weight
Q6_K
Q6_K
[2048, 4096]
blk.15.attn_q_norm.weight
F32
F32
[128]
blk.15.attn_sinks.weight
F32
F32
[32]
blk.15.attn_v.weight
Q6_K
Q6_K
[2048, 512]
blk.15.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.15.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.15.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.15.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.15.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.15.ffn_norm.weight
F32
F32
[2048]
blk.15.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.15.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.16
blk.16.attn_k.weight
Q6_K
Q6_K
[2048, 512]
blk.16.attn_k_norm.weight
F32
F32
[128]
blk.16.attn_norm.weight
F32
F32
[2048]
blk.16.attn_output.weight
Q6_K
Q6_K
[4096, 2048]
blk.16.attn_q.weight
Q6_K
Q6_K
[2048, 4096]
blk.16.attn_q_norm.weight
F32
F32
[128]
blk.16.attn_sinks.weight
F32
F32
[32]
blk.16.attn_v.weight
Q6_K
Q6_K
[2048, 512]
blk.16.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.16.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.16.ffn_gate_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.16.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.16.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.16.ffn_norm.weight
F32
F32
[2048]
blk.16.ffn_up_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.16.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.17
blk.17.attn_k.weight
Q8_0
Q8_0
[2048, 512]
blk.17.attn_k_norm.weight
F32
F32
[128]
blk.17.attn_norm.weight
F32
F32
[2048]
blk.17.attn_output.weight
Q8_0
Q8_0
[4096, 2048]
blk.17.attn_q.weight
Q8_0
Q8_0
[2048, 4096]
blk.17.attn_q_norm.weight
F32
F32
[128]
blk.17.attn_sinks.weight
F32
F32
[32]
blk.17.attn_v.weight
Q8_0
Q8_0
[2048, 512]
blk.17.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.17.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.17.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.17.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.17.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.17.ffn_norm.weight
F32
F32
[2048]
blk.17.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.17.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.18
blk.18.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.18.attn_k_norm.weight
F32
F32
[128]
blk.18.attn_norm.weight
F32
F32
[2048]
blk.18.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.18.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.18.attn_q_norm.weight
F32
F32
[128]
blk.18.attn_sinks.weight
F32
F32
[32]
blk.18.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.18.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.18.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.18.ffn_gate_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.18.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.18.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.18.ffn_norm.weight
F32
F32
[2048]
blk.18.ffn_up_exps.weight
IQ3_XXS
IQ3_XXS
[2048, 512, 256]
blk.18.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.19
blk.19.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.19.attn_k_norm.weight
F32
F32
[128]
blk.19.attn_norm.weight
F32
F32
[2048]
blk.19.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.19.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.19.attn_q_norm.weight
F32
F32
[128]
blk.19.attn_sinks.weight
F32
F32
[32]
blk.19.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.19.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.19.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.19.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.19.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.19.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.19.ffn_norm.weight
F32
F32
[2048]
blk.19.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.19.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.20
blk.20.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.20.attn_k_norm.weight
F32
F32
[128]
blk.20.attn_norm.weight
F32
F32
[2048]
blk.20.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.20.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.20.attn_q_norm.weight
F32
F32
[128]
blk.20.attn_sinks.weight
F32
F32
[32]
blk.20.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.20.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.20.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.20.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.20.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.20.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.20.ffn_norm.weight
F32
F32
[2048]
blk.20.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.20.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.21
blk.21.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.21.attn_k_norm.weight
F32
F32
[128]
blk.21.attn_norm.weight
F32
F32
[2048]
blk.21.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.21.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.21.attn_q_norm.weight
F32
F32
[128]
blk.21.attn_sinks.weight
F32
F32
[32]
blk.21.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.21.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.21.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.21.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.21.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.21.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.21.ffn_norm.weight
F32
F32
[2048]
blk.21.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.21.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.22
blk.22.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.22.attn_k_norm.weight
F32
F32
[128]
blk.22.attn_norm.weight
F32
F32
[2048]
blk.22.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.22.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.22.attn_q_norm.weight
F32
F32
[128]
blk.22.attn_sinks.weight
F32
F32
[32]
blk.22.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.22.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.22.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.22.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.22.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.22.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.22.ffn_norm.weight
F32
F32
[2048]
blk.22.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.22.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.23
blk.23.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.23.attn_k_norm.weight
F32
F32
[128]
blk.23.attn_norm.weight
F32
F32
[2048]
blk.23.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.23.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.23.attn_q_norm.weight
F32
F32
[128]
blk.23.attn_sinks.weight
F32
F32
[32]
blk.23.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.23.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.23.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.23.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.23.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.23.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.23.ffn_norm.weight
F32
F32
[2048]
blk.23.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.23.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.24
blk.24.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.24.attn_k_norm.weight
F32
F32
[128]
blk.24.attn_norm.weight
F32
F32
[2048]
blk.24.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.24.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.24.attn_q_norm.weight
F32
F32
[128]
blk.24.attn_sinks.weight
F32
F32
[32]
blk.24.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.24.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.24.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.24.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.24.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.24.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.24.ffn_norm.weight
F32
F32
[2048]
blk.24.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.24.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.25
blk.25.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.25.attn_k_norm.weight
F32
F32
[128]
blk.25.attn_norm.weight
F32
F32
[2048]
blk.25.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.25.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.25.attn_q_norm.weight
F32
F32
[128]
blk.25.attn_sinks.weight
F32
F32
[32]
blk.25.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.25.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.25.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.25.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.25.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.25.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.25.ffn_norm.weight
F32
F32
[2048]
blk.25.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.25.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.26
blk.26.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.26.attn_k_norm.weight
F32
F32
[128]
blk.26.attn_norm.weight
F32
F32
[2048]
blk.26.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.26.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.26.attn_q_norm.weight
F32
F32
[128]
blk.26.attn_sinks.weight
F32
F32
[32]
blk.26.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.26.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.26.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.26.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.26.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.26.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.26.ffn_norm.weight
F32
F32
[2048]
blk.26.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.26.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.27
blk.27.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.27.attn_k_norm.weight
F32
F32
[128]
blk.27.attn_norm.weight
F32
F32
[2048]
blk.27.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.27.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.27.attn_q_norm.weight
F32
F32
[128]
blk.27.attn_sinks.weight
F32
F32
[32]
blk.27.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.27.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.27.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.27.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.27.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.27.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.27.ffn_norm.weight
F32
F32
[2048]
blk.27.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.27.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.28
blk.28.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.28.attn_k_norm.weight
F32
F32
[128]
blk.28.attn_norm.weight
F32
F32
[2048]
blk.28.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.28.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.28.attn_q_norm.weight
F32
F32
[128]
blk.28.attn_sinks.weight
F32
F32
[32]
blk.28.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.28.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.28.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.28.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.28.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.28.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.28.ffn_norm.weight
F32
F32
[2048]
blk.28.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.28.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.29
blk.29.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.29.attn_k_norm.weight
F32
F32
[128]
blk.29.attn_norm.weight
F32
F32
[2048]
blk.29.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.29.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.29.attn_q_norm.weight
F32
F32
[128]
blk.29.attn_sinks.weight
F32
F32
[32]
blk.29.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.29.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.29.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.29.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.29.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.29.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.29.ffn_norm.weight
F32
F32
[2048]
blk.29.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.29.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.30
blk.30.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.30.attn_k_norm.weight
F32
F32
[128]
blk.30.attn_norm.weight
F32
F32
[2048]
blk.30.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.30.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.30.attn_q_norm.weight
F32
F32
[128]
blk.30.attn_sinks.weight
F32
F32
[32]
blk.30.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.30.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.30.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.30.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.30.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.30.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.30.ffn_norm.weight
F32
F32
[2048]
blk.30.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.30.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.31
blk.31.attn_k.weight
Q5_K
Q5_K
[2048, 512]
blk.31.attn_k_norm.weight
F32
F32
[128]
blk.31.attn_norm.weight
F32
F32
[2048]
blk.31.attn_output.weight
Q5_K
Q5_K
[4096, 2048]
blk.31.attn_q.weight
Q5_K
Q5_K
[2048, 4096]
blk.31.attn_q_norm.weight
F32
F32
[128]
blk.31.attn_sinks.weight
F32
F32
[32]
blk.31.attn_v.weight
Q5_K
Q5_K
[2048, 512]
blk.31.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.31.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.31.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.31.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.31.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.31.ffn_norm.weight
F32
F32
[2048]
blk.31.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.31.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.32
blk.32.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.32.attn_k_norm.weight
F32
F32
[128]
blk.32.attn_norm.weight
F32
F32
[2048]
blk.32.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.32.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.32.attn_q_norm.weight
F32
F32
[128]
blk.32.attn_sinks.weight
F32
F32
[32]
blk.32.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.32.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.32.ffn_down_shexp.weight
Q4_K
Q4_K
[1024, 2048]
blk.32.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.32.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.32.ffn_gate_shexp.weight
Q4_K
Q4_K
[2048, 1024]
blk.32.ffn_norm.weight
F32
F32
[2048]
blk.32.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.32.ffn_up_shexp.weight
Q4_K
Q4_K
[2048, 1024]
blk.33
blk.33.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.33.attn_k_norm.weight
F32
F32
[128]
blk.33.attn_norm.weight
F32
F32
[2048]
blk.33.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.33.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.33.attn_q_norm.weight
F32
F32
[128]
blk.33.attn_sinks.weight
F32
F32
[32]
blk.33.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.33.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.33.ffn_down_shexp.weight
IQ4_XS
IQ4_XS
[1024, 2048]
blk.33.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.33.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.33.ffn_gate_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.33.ffn_norm.weight
F32
F32
[2048]
blk.33.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.33.ffn_up_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.34
blk.34.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.34.attn_k_norm.weight
F32
F32
[128]
blk.34.attn_norm.weight
F32
F32
[2048]
blk.34.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.34.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.34.attn_q_norm.weight
F32
F32
[128]
blk.34.attn_sinks.weight
F32
F32
[32]
blk.34.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.34.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.34.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.34.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.34.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.34.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.34.ffn_norm.weight
F32
F32
[2048]
blk.34.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.34.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.35
blk.35.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.35.attn_k_norm.weight
F32
F32
[128]
blk.35.attn_norm.weight
F32
F32
[2048]
blk.35.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.35.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.35.attn_q_norm.weight
F32
F32
[128]
blk.35.attn_sinks.weight
F32
F32
[32]
blk.35.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.35.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.35.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.35.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.35.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.35.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.35.ffn_norm.weight
F32
F32
[2048]
blk.35.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.35.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.36
blk.36.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.36.attn_k_norm.weight
F32
F32
[128]
blk.36.attn_norm.weight
F32
F32
[2048]
blk.36.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.36.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.36.attn_q_norm.weight
F32
F32
[128]
blk.36.attn_sinks.weight
F32
F32
[32]
blk.36.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.36.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.36.ffn_down_shexp.weight
IQ3_XXS
IQ3_XXS
[1024, 2048]
blk.36.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.36.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.36.ffn_gate_shexp.weight
IQ3_XXS
IQ3_XXS
[2048, 1024]
blk.36.ffn_norm.weight
F32
F32
[2048]
blk.36.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.36.ffn_up_shexp.weight
IQ3_XXS
IQ3_XXS
[2048, 1024]
blk.37
blk.37.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.37.attn_k_norm.weight
F32
F32
[128]
blk.37.attn_norm.weight
F32
F32
[2048]
blk.37.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.37.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.37.attn_q_norm.weight
F32
F32
[128]
blk.37.attn_sinks.weight
F32
F32
[32]
blk.37.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.37.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.37.ffn_down_shexp.weight
Q4_K
Q4_K
[1024, 2048]
blk.37.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.37.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.37.ffn_gate_shexp.weight
Q4_K
Q4_K
[2048, 1024]
blk.37.ffn_norm.weight
F32
F32
[2048]
blk.37.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.37.ffn_up_shexp.weight
Q4_K
Q4_K
[2048, 1024]
blk.38
blk.38.attn_k.weight
IQ3_S
IQ3_S
[2048, 512]
blk.38.attn_k_norm.weight
F32
F32
[128]
blk.38.attn_norm.weight
F32
F32
[2048]
blk.38.attn_output.weight
IQ3_S
IQ3_S
[4096, 2048]
blk.38.attn_q.weight
IQ3_S
IQ3_S
[2048, 4096]
blk.38.attn_q_norm.weight
F32
F32
[128]
blk.38.attn_sinks.weight
F32
F32
[32]
blk.38.attn_v.weight
IQ3_S
IQ3_S
[2048, 512]
blk.38.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.38.ffn_down_shexp.weight
IQ4_XS
IQ4_XS
[1024, 2048]
blk.38.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.38.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.38.ffn_gate_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.38.ffn_norm.weight
F32
F32
[2048]
blk.38.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.38.ffn_up_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.39
blk.39.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.39.attn_k_norm.weight
F32
F32
[128]
blk.39.attn_norm.weight
F32
F32
[2048]
blk.39.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.39.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.39.attn_q_norm.weight
F32
F32
[128]
blk.39.attn_sinks.weight
F32
F32
[32]
blk.39.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.39.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.39.ffn_down_shexp.weight
IQ4_XS
IQ4_XS
[1024, 2048]
blk.39.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.39.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.39.ffn_gate_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.39.ffn_norm.weight
F32
F32
[2048]
blk.39.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.39.ffn_up_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.40
blk.40.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.40.attn_k_norm.weight
F32
F32
[128]
blk.40.attn_norm.weight
F32
F32
[2048]
blk.40.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.40.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.40.attn_q_norm.weight
F32
F32
[128]
blk.40.attn_sinks.weight
F32
F32
[32]
blk.40.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.40.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.40.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.40.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.40.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.40.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.40.ffn_norm.weight
F32
F32
[2048]
blk.40.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.40.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.41
blk.41.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.41.attn_k_norm.weight
F32
F32
[128]
blk.41.attn_norm.weight
F32
F32
[2048]
blk.41.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.41.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.41.attn_q_norm.weight
F32
F32
[128]
blk.41.attn_sinks.weight
F32
F32
[32]
blk.41.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.41.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.41.ffn_down_shexp.weight
IQ4_XS
IQ4_XS
[1024, 2048]
blk.41.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.41.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.41.ffn_gate_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.41.ffn_norm.weight
F32
F32
[2048]
blk.41.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.41.ffn_up_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.42
blk.42.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.42.attn_k_norm.weight
F32
F32
[128]
blk.42.attn_norm.weight
F32
F32
[2048]
blk.42.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.42.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.42.attn_q_norm.weight
F32
F32
[128]
blk.42.attn_sinks.weight
F32
F32
[32]
blk.42.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.42.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.42.ffn_down_shexp.weight
IQ4_XS
IQ4_XS
[1024, 2048]
blk.42.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.42.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.42.ffn_gate_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.42.ffn_norm.weight
F32
F32
[2048]
blk.42.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.42.ffn_up_shexp.weight
IQ4_XS
IQ4_XS
[2048, 1024]
blk.43
blk.43.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.43.attn_k_norm.weight
F32
F32
[128]
blk.43.attn_norm.weight
F32
F32
[2048]
blk.43.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.43.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.43.attn_q_norm.weight
F32
F32
[128]
blk.43.attn_sinks.weight
F32
F32
[32]
blk.43.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.43.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.43.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.43.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.43.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.43.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.43.ffn_norm.weight
F32
F32
[2048]
blk.43.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.43.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.44
blk.44.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.44.attn_k_norm.weight
F32
F32
[128]
blk.44.attn_norm.weight
F32
F32
[2048]
blk.44.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.44.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.44.attn_q_norm.weight
F32
F32
[128]
blk.44.attn_sinks.weight
F32
F32
[32]
blk.44.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.44.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.44.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.44.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.44.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.44.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.44.ffn_norm.weight
F32
F32
[2048]
blk.44.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.44.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.45
blk.45.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.45.attn_k_norm.weight
F32
F32
[128]
blk.45.attn_norm.weight
F32
F32
[2048]
blk.45.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.45.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.45.attn_q_norm.weight
F32
F32
[128]
blk.45.attn_sinks.weight
F32
F32
[32]
blk.45.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.45.ffn_down_exps.weight
IQ2_XXS
IQ2_XXS
[512, 2048, 256]
blk.45.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.45.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.45.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.45.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.45.ffn_norm.weight
F32
F32
[2048]
blk.45.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.45.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.46
blk.46.attn_k.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.46.attn_k_norm.weight
F32
F32
[128]
blk.46.attn_norm.weight
F32
F32
[2048]
blk.46.attn_output.weight
IQ4_XS
IQ4_XS
[4096, 2048]
blk.46.attn_q.weight
IQ4_XS
IQ4_XS
[2048, 4096]
blk.46.attn_q_norm.weight
F32
F32
[128]
blk.46.attn_sinks.weight
F32
F32
[32]
blk.46.attn_v.weight
IQ4_XS
IQ4_XS
[2048, 512]
blk.46.ffn_down_exps.weight
IQ3_XXS
IQ3_XXS
[512, 2048, 256]
blk.46.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.46.ffn_gate_exps.weight
Q2_K
Q2_K
[2048, 512, 256]
blk.46.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.46.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.46.ffn_norm.weight
F32
F32
[2048]
blk.46.ffn_up_exps.weight
Q2_K
Q2_K
[2048, 512, 256]
blk.46.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.47
blk.47.attn_k.weight
IQ3_S
IQ3_S
[2048, 512]
blk.47.attn_k_norm.weight
F32
F32
[128]
blk.47.attn_norm.weight
F32
F32
[2048]
blk.47.attn_output.weight
IQ3_S
IQ3_S
[4096, 2048]
blk.47.attn_q.weight
IQ3_S
IQ3_S
[2048, 4096]
blk.47.attn_q_norm.weight
F32
F32
[128]
blk.47.attn_sinks.weight
F32
F32
[32]
blk.47.attn_v.weight
IQ3_S
IQ3_S
[2048, 512]
blk.47.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.47.ffn_down_shexp.weight
Q5_K
Q5_K
[1024, 2048]
blk.47.ffn_gate_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.47.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.47.ffn_gate_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.47.ffn_norm.weight
F32
F32
[2048]
blk.47.ffn_up_exps.weight
IQ2_XXS
IQ2_XXS
[2048, 512, 256]
blk.47.ffn_up_shexp.weight
Q5_K
Q5_K
[2048, 1024]
blk.48
blk.48.attn_k.weight
IQ3_XXS
IQ3_XXS
[2048, 512]
blk.48.attn_k_norm.weight
F32
F32
[128]
blk.48.attn_norm.weight
F32
F32
[2048]
blk.48.attn_output.weight
IQ3_XXS
IQ3_XXS
[4096, 2048]
blk.48.attn_q.weight
IQ3_XXS
IQ3_XXS
[2048, 4096]
blk.48.attn_q_norm.weight
F32
F32
[128]
blk.48.attn_sinks.weight
F32
F32
[32]
blk.48.attn_v.weight
IQ3_XXS
IQ3_XXS
[2048, 512]
blk.48.ffn_down_exps.weight
IQ2_S
IQ2_S
[512, 2048, 256]
blk.48.ffn_down_shexp.weight
Q8_0
Q8_0
[1024, 2048]
blk.48.ffn_gate_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.48.ffn_gate_inp.weight
BF16
BF16
[2048, 256]
blk.48.ffn_gate_shexp.weight
Q8_0
Q8_0
[2048, 1024]
blk.48.ffn_norm.weight
F32
F32
[2048]
blk.48.ffn_up_exps.weight
IQ2_S
IQ2_S
[2048, 512, 256]
blk.48.ffn_up_shexp.weight
Q8_0
Q8_0
[2048, 1024]
output.weight
Q6_K
Q6_K
[2048, 150144]
output_norm.weight
F32
F32
[2048]