17 2 weeks ago

This is an AI model based on the Gemma3 Model Family, but designed to address issues such as sycophancy or unnecessary praise that may arise in AI models. This has no connection to the OpenAssistans project; this AI model was created by OpenSoftware-World

vision 270m 1b 4b 12b 27b
7cd4618c1faf · 815MB
    Metadata
  • general.architecture
    gemma3
  • general.file_type
    Q4_K_M
  • gemma3.attention.head_count
    4
  • gemma3.attention.head_count_kv
    1
  • gemma3.attention.key_length
    256
  • gemma3.attention.layer_norm_rms_epsilon
    1e-06
  • gemma3.attention.sliding_window
    512
  • gemma3.attention.value_length
    256
  • gemma3.block_count
    26
  • gemma3.context_length
    32768
  • gemma3.embedding_length
    1152
  • gemma3.feed_forward_length
    6912
  • gemma3.final_logit_softcapping
    30
  • gemma3.rope.global.freq_base
    1e+06
  • gemma3.rope.local.freq_base
    10000
  • tokenizer.ggml.add_bos_token
    true
  • tokenizer.ggml.add_eos_token
    false
  • tokenizer.ggml.add_padding_token
    false
  • tokenizer.ggml.add_unknown_token
    false
  • tokenizer.ggml.bos_token_id
    2
  • tokenizer.ggml.eos_token_id
    1
  • tokenizer.ggml.merges
    [ , ▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ ▁, , , ▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁ ▁▁, ...]
  • tokenizer.ggml.model
    llama
  • tokenizer.ggml.padding_token_id
    0
  • tokenizer.ggml.pre
    default
  • tokenizer.ggml.scores
    [0, 0, 0, 0, 0, ...]
  • tokenizer.ggml.token_type
    [3, 3, 3, 2, 1, ...]
  • tokenizer.ggml.tokens
    [<pad>, <eos>, <bos>, <unk>, <mask>, ...]
  • tokenizer.ggml.unknown_token_id
    3
  • Tensor
  • token_embd.weight
    Q8_0
    [1152, 262144]
  • blk.0
  • blk.0.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.0.attn_k_norm.weight
    F32
    [256]
  • blk.0.attn_norm.weight
    F32
    [1152]
  • blk.0.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.0.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.0.attn_q_norm.weight
    F32
    [256]
  • blk.0.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.0.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.0.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.0.ffn_norm.weight
    F32
    [1152]
  • blk.0.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.0.post_attention_norm.weight
    F32
    [1152]
  • blk.0.post_ffw_norm.weight
    F32
    [1152]
  • blk.1
  • blk.1.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.1.attn_k_norm.weight
    F32
    [256]
  • blk.1.attn_norm.weight
    F32
    [1152]
  • blk.1.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.1.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.1.attn_q_norm.weight
    F32
    [256]
  • blk.1.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.1.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.1.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.1.ffn_norm.weight
    F32
    [1152]
  • blk.1.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.1.post_attention_norm.weight
    F32
    [1152]
  • blk.1.post_ffw_norm.weight
    F32
    [1152]
  • blk.2
  • blk.2.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.2.attn_k_norm.weight
    F32
    [256]
  • blk.2.attn_norm.weight
    F32
    [1152]
  • blk.2.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.2.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.2.attn_q_norm.weight
    F32
    [256]
  • blk.2.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.2.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.2.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.2.ffn_norm.weight
    F32
    [1152]
  • blk.2.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.2.post_attention_norm.weight
    F32
    [1152]
  • blk.2.post_ffw_norm.weight
    F32
    [1152]
  • blk.3
  • blk.3.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.3.attn_k_norm.weight
    F32
    [256]
  • blk.3.attn_norm.weight
    F32
    [1152]
  • blk.3.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.3.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.3.attn_q_norm.weight
    F32
    [256]
  • blk.3.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.3.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.3.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.3.ffn_norm.weight
    F32
    [1152]
  • blk.3.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.3.post_attention_norm.weight
    F32
    [1152]
  • blk.3.post_ffw_norm.weight
    F32
    [1152]
  • blk.4
  • blk.4.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.4.attn_k_norm.weight
    F32
    [256]
  • blk.4.attn_norm.weight
    F32
    [1152]
  • blk.4.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.4.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.4.attn_q_norm.weight
    F32
    [256]
  • blk.4.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.4.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.4.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.4.ffn_norm.weight
    F32
    [1152]
  • blk.4.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.4.post_attention_norm.weight
    F32
    [1152]
  • blk.4.post_ffw_norm.weight
    F32
    [1152]
  • blk.5
  • blk.5.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.5.attn_k_norm.weight
    F32
    [256]
  • blk.5.attn_norm.weight
    F32
    [1152]
  • blk.5.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.5.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.5.attn_q_norm.weight
    F32
    [256]
  • blk.5.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.5.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.5.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.5.ffn_norm.weight
    F32
    [1152]
  • blk.5.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.5.post_attention_norm.weight
    F32
    [1152]
  • blk.5.post_ffw_norm.weight
    F32
    [1152]
  • blk.6
  • blk.6.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.6.attn_k_norm.weight
    F32
    [256]
  • blk.6.attn_norm.weight
    F32
    [1152]
  • blk.6.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.6.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.6.attn_q_norm.weight
    F32
    [256]
  • blk.6.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.6.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.6.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.6.ffn_norm.weight
    F32
    [1152]
  • blk.6.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.6.post_attention_norm.weight
    F32
    [1152]
  • blk.6.post_ffw_norm.weight
    F32
    [1152]
  • blk.7
  • blk.7.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.7.attn_k_norm.weight
    F32
    [256]
  • blk.7.attn_norm.weight
    F32
    [1152]
  • blk.7.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.7.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.7.attn_q_norm.weight
    F32
    [256]
  • blk.7.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.7.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.7.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.7.ffn_norm.weight
    F32
    [1152]
  • blk.7.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.7.post_attention_norm.weight
    F32
    [1152]
  • blk.7.post_ffw_norm.weight
    F32
    [1152]
  • blk.8
  • blk.8.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.8.attn_k_norm.weight
    F32
    [256]
  • blk.8.attn_norm.weight
    F32
    [1152]
  • blk.8.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.8.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.8.attn_q_norm.weight
    F32
    [256]
  • blk.8.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.8.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.8.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.8.ffn_norm.weight
    F32
    [1152]
  • blk.8.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.8.post_attention_norm.weight
    F32
    [1152]
  • blk.8.post_ffw_norm.weight
    F32
    [1152]
  • blk.9
  • blk.9.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.9.attn_k_norm.weight
    F32
    [256]
  • blk.9.attn_norm.weight
    F32
    [1152]
  • blk.9.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.9.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.9.attn_q_norm.weight
    F32
    [256]
  • blk.9.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.9.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.9.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.9.ffn_norm.weight
    F32
    [1152]
  • blk.9.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.9.post_attention_norm.weight
    F32
    [1152]
  • blk.9.post_ffw_norm.weight
    F32
    [1152]
  • blk.10
  • blk.10.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.10.attn_k_norm.weight
    F32
    [256]
  • blk.10.attn_norm.weight
    F32
    [1152]
  • blk.10.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.10.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.10.attn_q_norm.weight
    F32
    [256]
  • blk.10.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.10.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.10.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.10.ffn_norm.weight
    F32
    [1152]
  • blk.10.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.10.post_attention_norm.weight
    F32
    [1152]
  • blk.10.post_ffw_norm.weight
    F32
    [1152]
  • blk.11
  • blk.11.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.11.attn_k_norm.weight
    F32
    [256]
  • blk.11.attn_norm.weight
    F32
    [1152]
  • blk.11.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.11.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.11.attn_q_norm.weight
    F32
    [256]
  • blk.11.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.11.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.11.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.11.ffn_norm.weight
    F32
    [1152]
  • blk.11.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.11.post_attention_norm.weight
    F32
    [1152]
  • blk.11.post_ffw_norm.weight
    F32
    [1152]
  • blk.12
  • blk.12.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.12.attn_k_norm.weight
    F32
    [256]
  • blk.12.attn_norm.weight
    F32
    [1152]
  • blk.12.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.12.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.12.attn_q_norm.weight
    F32
    [256]
  • blk.12.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.12.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.12.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.12.ffn_norm.weight
    F32
    [1152]
  • blk.12.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.12.post_attention_norm.weight
    F32
    [1152]
  • blk.12.post_ffw_norm.weight
    F32
    [1152]
  • blk.13
  • blk.13.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.13.attn_k_norm.weight
    F32
    [256]
  • blk.13.attn_norm.weight
    F32
    [1152]
  • blk.13.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.13.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.13.attn_q_norm.weight
    F32
    [256]
  • blk.13.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.13.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.13.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.13.ffn_norm.weight
    F32
    [1152]
  • blk.13.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.13.post_attention_norm.weight
    F32
    [1152]
  • blk.13.post_ffw_norm.weight
    F32
    [1152]
  • blk.14
  • blk.14.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.14.attn_k_norm.weight
    F32
    [256]
  • blk.14.attn_norm.weight
    F32
    [1152]
  • blk.14.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.14.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.14.attn_q_norm.weight
    F32
    [256]
  • blk.14.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.14.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.14.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.14.ffn_norm.weight
    F32
    [1152]
  • blk.14.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.14.post_attention_norm.weight
    F32
    [1152]
  • blk.14.post_ffw_norm.weight
    F32
    [1152]
  • blk.15
  • blk.15.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.15.attn_k_norm.weight
    F32
    [256]
  • blk.15.attn_norm.weight
    F32
    [1152]
  • blk.15.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.15.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.15.attn_q_norm.weight
    F32
    [256]
  • blk.15.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.15.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.15.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.15.ffn_norm.weight
    F32
    [1152]
  • blk.15.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.15.post_attention_norm.weight
    F32
    [1152]
  • blk.15.post_ffw_norm.weight
    F32
    [1152]
  • blk.16
  • blk.16.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.16.attn_k_norm.weight
    F32
    [256]
  • blk.16.attn_norm.weight
    F32
    [1152]
  • blk.16.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.16.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.16.attn_q_norm.weight
    F32
    [256]
  • blk.16.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.16.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.16.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.16.ffn_norm.weight
    F32
    [1152]
  • blk.16.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.16.post_attention_norm.weight
    F32
    [1152]
  • blk.16.post_ffw_norm.weight
    F32
    [1152]
  • blk.17
  • blk.17.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.17.attn_k_norm.weight
    F32
    [256]
  • blk.17.attn_norm.weight
    F32
    [1152]
  • blk.17.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.17.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.17.attn_q_norm.weight
    F32
    [256]
  • blk.17.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.17.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.17.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.17.ffn_norm.weight
    F32
    [1152]
  • blk.17.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.17.post_attention_norm.weight
    F32
    [1152]
  • blk.17.post_ffw_norm.weight
    F32
    [1152]
  • blk.18
  • blk.18.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.18.attn_k_norm.weight
    F32
    [256]
  • blk.18.attn_norm.weight
    F32
    [1152]
  • blk.18.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.18.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.18.attn_q_norm.weight
    F32
    [256]
  • blk.18.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.18.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.18.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.18.ffn_norm.weight
    F32
    [1152]
  • blk.18.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.18.post_attention_norm.weight
    F32
    [1152]
  • blk.18.post_ffw_norm.weight
    F32
    [1152]
  • blk.19
  • blk.19.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.19.attn_k_norm.weight
    F32
    [256]
  • blk.19.attn_norm.weight
    F32
    [1152]
  • blk.19.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.19.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.19.attn_q_norm.weight
    F32
    [256]
  • blk.19.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.19.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.19.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.19.ffn_norm.weight
    F32
    [1152]
  • blk.19.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.19.post_attention_norm.weight
    F32
    [1152]
  • blk.19.post_ffw_norm.weight
    F32
    [1152]
  • blk.20
  • blk.20.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.20.attn_k_norm.weight
    F32
    [256]
  • blk.20.attn_norm.weight
    F32
    [1152]
  • blk.20.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.20.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.20.attn_q_norm.weight
    F32
    [256]
  • blk.20.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.20.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.20.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.20.ffn_norm.weight
    F32
    [1152]
  • blk.20.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.20.post_attention_norm.weight
    F32
    [1152]
  • blk.20.post_ffw_norm.weight
    F32
    [1152]
  • blk.21
  • blk.21.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.21.attn_k_norm.weight
    F32
    [256]
  • blk.21.attn_norm.weight
    F32
    [1152]
  • blk.21.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.21.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.21.attn_q_norm.weight
    F32
    [256]
  • blk.21.attn_v.weight
    Q5_0
    [1152, 256]
  • blk.21.ffn_down.weight
    Q4_K
    [6912, 1152]
  • blk.21.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.21.ffn_norm.weight
    F32
    [1152]
  • blk.21.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.21.post_attention_norm.weight
    F32
    [1152]
  • blk.21.post_ffw_norm.weight
    F32
    [1152]
  • blk.22
  • blk.22.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.22.attn_k_norm.weight
    F32
    [256]
  • blk.22.attn_norm.weight
    F32
    [1152]
  • blk.22.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.22.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.22.attn_q_norm.weight
    F32
    [256]
  • blk.22.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.22.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.22.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.22.ffn_norm.weight
    F32
    [1152]
  • blk.22.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.22.post_attention_norm.weight
    F32
    [1152]
  • blk.22.post_ffw_norm.weight
    F32
    [1152]
  • blk.23
  • blk.23.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.23.attn_k_norm.weight
    F32
    [256]
  • blk.23.attn_norm.weight
    F32
    [1152]
  • blk.23.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.23.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.23.attn_q_norm.weight
    F32
    [256]
  • blk.23.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.23.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.23.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.23.ffn_norm.weight
    F32
    [1152]
  • blk.23.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.23.post_attention_norm.weight
    F32
    [1152]
  • blk.23.post_ffw_norm.weight
    F32
    [1152]
  • blk.24
  • blk.24.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.24.attn_k_norm.weight
    F32
    [256]
  • blk.24.attn_norm.weight
    F32
    [1152]
  • blk.24.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.24.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.24.attn_q_norm.weight
    F32
    [256]
  • blk.24.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.24.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.24.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.24.ffn_norm.weight
    F32
    [1152]
  • blk.24.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.24.post_attention_norm.weight
    F32
    [1152]
  • blk.24.post_ffw_norm.weight
    F32
    [1152]
  • blk.25
  • blk.25.attn_k.weight
    Q5_0
    [1152, 256]
  • blk.25.attn_k_norm.weight
    F32
    [256]
  • blk.25.attn_norm.weight
    F32
    [1152]
  • blk.25.attn_output.weight
    Q4_K
    [1024, 1152]
  • blk.25.attn_q.weight
    Q5_0
    [1152, 1024]
  • blk.25.attn_q_norm.weight
    F32
    [256]
  • blk.25.attn_v.weight
    Q8_0
    [1152, 256]
  • blk.25.ffn_down.weight
    Q6_K
    [6912, 1152]
  • blk.25.ffn_gate.weight
    Q5_0
    [1152, 6912]
  • blk.25.ffn_norm.weight
    F32
    [1152]
  • blk.25.ffn_up.weight
    Q5_0
    [1152, 6912]
  • blk.25.post_attention_norm.weight
    F32
    [1152]
  • blk.25.post_ffw_norm.weight
    F32
    [1152]
  • output_norm.weight
    F32
    [1152]