alfred

A robust conversational model designed to be used for both chat and instruct use cases.

40b

14K 13 months ago

7 Tags
... /
model
cc9f2665d100 · 24GB
    Metadata
  • falcon.attention.head_count
    128
  • falcon.attention.head_count_kv
    8
  • falcon.attention.layer_norm_epsilon
    1e-05
  • falcon.block_count
    60
  • falcon.context_length
    2048
  • falcon.embedding_length
    8192
  • falcon.feed_forward_length
    32768
  • falcon.tensor_data_layout
    jploski
  • general.architecture
    falcon
  • general.file_type
    2
  • general.name
    Falcon
  • general.quantization_version
    2
  • tokenizer.ggml.eos_token_id
    65020
  • tokenizer.ggml.merges
    [Ġ t, Ġ a, i n, h e, r e, ...]
  • tokenizer.ggml.model
    gpt2
  • tokenizer.ggml.scores
    [0, 0, 0, 0, 0, ...]
  • tokenizer.ggml.token_type
    [1, 1, 1, 1, 1, ...]
  • tokenizer.ggml.tokens
    [>>TITLE<<, >>ABSTRACT<<, >>INTRODUCTION<<, >>SUMMARY<<, >>COMMENT<<, ...]
    • Tensor
  • Name
    Type
    Shape
  • token_embd.weight
    Q4_0
    [8192, 65024]
    • blk.0
  • blk.0.attn_norm.bias
    F32
    [8192]
  • blk.0.attn_norm.weight
    F32
    [8192]
  • blk.0.attn_norm_2.bias
    F32
    [8192]
  • blk.0.attn_norm_2.weight
    F32
    [8192]
  • blk.0.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.0.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.0.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.0.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.1
  • blk.1.attn_norm.bias
    F32
    [8192]
  • blk.1.attn_norm.weight
    F32
    [8192]
  • blk.1.attn_norm_2.bias
    F32
    [8192]
  • blk.1.attn_norm_2.weight
    F32
    [8192]
  • blk.1.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.1.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.1.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.1.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.2
  • blk.2.attn_norm.bias
    F32
    [8192]
  • blk.2.attn_norm.weight
    F32
    [8192]
  • blk.2.attn_norm_2.bias
    F32
    [8192]
  • blk.2.attn_norm_2.weight
    F32
    [8192]
  • blk.2.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.2.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.2.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.2.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.3
  • blk.3.attn_norm.bias
    F32
    [8192]
  • blk.3.attn_norm.weight
    F32
    [8192]
  • blk.3.attn_norm_2.bias
    F32
    [8192]
  • blk.3.attn_norm_2.weight
    F32
    [8192]
  • blk.3.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.3.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.3.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.3.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.4
  • blk.4.attn_norm.bias
    F32
    [8192]
  • blk.4.attn_norm.weight
    F32
    [8192]
  • blk.4.attn_norm_2.bias
    F32
    [8192]
  • blk.4.attn_norm_2.weight
    F32
    [8192]
  • blk.4.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.4.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.4.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.4.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.5
  • blk.5.attn_norm.bias
    F32
    [8192]
  • blk.5.attn_norm.weight
    F32
    [8192]
  • blk.5.attn_norm_2.bias
    F32
    [8192]
  • blk.5.attn_norm_2.weight
    F32
    [8192]
  • blk.5.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.5.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.5.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.5.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.6
  • blk.6.attn_norm.bias
    F32
    [8192]
  • blk.6.attn_norm.weight
    F32
    [8192]
  • blk.6.attn_norm_2.bias
    F32
    [8192]
  • blk.6.attn_norm_2.weight
    F32
    [8192]
  • blk.6.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.6.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.6.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.6.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.7
  • blk.7.attn_norm.bias
    F32
    [8192]
  • blk.7.attn_norm.weight
    F32
    [8192]
  • blk.7.attn_norm_2.bias
    F32
    [8192]
  • blk.7.attn_norm_2.weight
    F32
    [8192]
  • blk.7.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.7.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.7.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.7.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.8
  • blk.8.attn_norm.bias
    F32
    [8192]
  • blk.8.attn_norm.weight
    F32
    [8192]
  • blk.8.attn_norm_2.bias
    F32
    [8192]
  • blk.8.attn_norm_2.weight
    F32
    [8192]
  • blk.8.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.8.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.8.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.8.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.9
  • blk.9.attn_norm.bias
    F32
    [8192]
  • blk.9.attn_norm.weight
    F32
    [8192]
  • blk.9.attn_norm_2.bias
    F32
    [8192]
  • blk.9.attn_norm_2.weight
    F32
    [8192]
  • blk.9.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.9.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.9.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.9.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.10
  • blk.10.attn_norm.bias
    F32
    [8192]
  • blk.10.attn_norm.weight
    F32
    [8192]
  • blk.10.attn_norm_2.bias
    F32
    [8192]
  • blk.10.attn_norm_2.weight
    F32
    [8192]
  • blk.10.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.10.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.10.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.10.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.11
  • blk.11.attn_norm.bias
    F32
    [8192]
  • blk.11.attn_norm.weight
    F32
    [8192]
  • blk.11.attn_norm_2.bias
    F32
    [8192]
  • blk.11.attn_norm_2.weight
    F32
    [8192]
  • blk.11.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.11.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.11.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.11.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.12
  • blk.12.attn_norm.bias
    F32
    [8192]
  • blk.12.attn_norm.weight
    F32
    [8192]
  • blk.12.attn_norm_2.bias
    F32
    [8192]
  • blk.12.attn_norm_2.weight
    F32
    [8192]
  • blk.12.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.12.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.12.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.12.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.13
  • blk.13.attn_norm.bias
    F32
    [8192]
  • blk.13.attn_norm.weight
    F32
    [8192]
  • blk.13.attn_norm_2.bias
    F32
    [8192]
  • blk.13.attn_norm_2.weight
    F32
    [8192]
  • blk.13.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.13.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.13.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.13.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.14
  • blk.14.attn_norm.bias
    F32
    [8192]
  • blk.14.attn_norm.weight
    F32
    [8192]
  • blk.14.attn_norm_2.bias
    F32
    [8192]
  • blk.14.attn_norm_2.weight
    F32
    [8192]
  • blk.14.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.14.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.14.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.14.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.15
  • blk.15.attn_norm.bias
    F32
    [8192]
  • blk.15.attn_norm.weight
    F32
    [8192]
  • blk.15.attn_norm_2.bias
    F32
    [8192]
  • blk.15.attn_norm_2.weight
    F32
    [8192]
  • blk.15.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.15.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.15.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.15.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.16
  • blk.16.attn_norm.bias
    F32
    [8192]
  • blk.16.attn_norm.weight
    F32
    [8192]
  • blk.16.attn_norm_2.bias
    F32
    [8192]
  • blk.16.attn_norm_2.weight
    F32
    [8192]
  • blk.16.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.16.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.16.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.16.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.17
  • blk.17.attn_norm.bias
    F32
    [8192]
  • blk.17.attn_norm.weight
    F32
    [8192]
  • blk.17.attn_norm_2.bias
    F32
    [8192]
  • blk.17.attn_norm_2.weight
    F32
    [8192]
  • blk.17.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.17.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.17.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.17.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.18
  • blk.18.attn_norm.bias
    F32
    [8192]
  • blk.18.attn_norm.weight
    F32
    [8192]
  • blk.18.attn_norm_2.bias
    F32
    [8192]
  • blk.18.attn_norm_2.weight
    F32
    [8192]
  • blk.18.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.18.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.18.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.18.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.19
  • blk.19.attn_norm.bias
    F32
    [8192]
  • blk.19.attn_norm.weight
    F32
    [8192]
  • blk.19.attn_norm_2.bias
    F32
    [8192]
  • blk.19.attn_norm_2.weight
    F32
    [8192]
  • blk.19.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.19.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.19.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.19.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.20
  • blk.20.attn_norm.bias
    F32
    [8192]
  • blk.20.attn_norm.weight
    F32
    [8192]
  • blk.20.attn_norm_2.bias
    F32
    [8192]
  • blk.20.attn_norm_2.weight
    F32
    [8192]
  • blk.20.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.20.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.20.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.20.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.21
  • blk.21.attn_norm.bias
    F32
    [8192]
  • blk.21.attn_norm.weight
    F32
    [8192]
  • blk.21.attn_norm_2.bias
    F32
    [8192]
  • blk.21.attn_norm_2.weight
    F32
    [8192]
  • blk.21.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.21.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.21.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.21.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.22
  • blk.22.attn_norm.bias
    F32
    [8192]
  • blk.22.attn_norm.weight
    F32
    [8192]
  • blk.22.attn_norm_2.bias
    F32
    [8192]
  • blk.22.attn_norm_2.weight
    F32
    [8192]
  • blk.22.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.22.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.22.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.22.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.23
  • blk.23.attn_norm.bias
    F32
    [8192]
  • blk.23.attn_norm.weight
    F32
    [8192]
  • blk.23.attn_norm_2.bias
    F32
    [8192]
  • blk.23.attn_norm_2.weight
    F32
    [8192]
  • blk.23.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.23.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.23.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.23.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.24
  • blk.24.attn_norm.bias
    F32
    [8192]
  • blk.24.attn_norm.weight
    F32
    [8192]
  • blk.24.attn_norm_2.bias
    F32
    [8192]
  • blk.24.attn_norm_2.weight
    F32
    [8192]
  • blk.24.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.24.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.24.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.24.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.25
  • blk.25.attn_norm.bias
    F32
    [8192]
  • blk.25.attn_norm.weight
    F32
    [8192]
  • blk.25.attn_norm_2.bias
    F32
    [8192]
  • blk.25.attn_norm_2.weight
    F32
    [8192]
  • blk.25.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.25.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.25.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.25.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.26
  • blk.26.attn_norm.bias
    F32
    [8192]
  • blk.26.attn_norm.weight
    F32
    [8192]
  • blk.26.attn_norm_2.bias
    F32
    [8192]
  • blk.26.attn_norm_2.weight
    F32
    [8192]
  • blk.26.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.26.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.26.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.26.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.27
  • blk.27.attn_norm.bias
    F32
    [8192]
  • blk.27.attn_norm.weight
    F32
    [8192]
  • blk.27.attn_norm_2.bias
    F32
    [8192]
  • blk.27.attn_norm_2.weight
    F32
    [8192]
  • blk.27.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.27.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.27.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.27.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.28
  • blk.28.attn_norm.bias
    F32
    [8192]
  • blk.28.attn_norm.weight
    F32
    [8192]
  • blk.28.attn_norm_2.bias
    F32
    [8192]
  • blk.28.attn_norm_2.weight
    F32
    [8192]
  • blk.28.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.28.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.28.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.28.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.29
  • blk.29.attn_norm.bias
    F32
    [8192]
  • blk.29.attn_norm.weight
    F32
    [8192]
  • blk.29.attn_norm_2.bias
    F32
    [8192]
  • blk.29.attn_norm_2.weight
    F32
    [8192]
  • blk.29.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.29.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.29.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.29.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.30
  • blk.30.attn_norm.bias
    F32
    [8192]
  • blk.30.attn_norm.weight
    F32
    [8192]
  • blk.30.attn_norm_2.bias
    F32
    [8192]
  • blk.30.attn_norm_2.weight
    F32
    [8192]
  • blk.30.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.30.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.30.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.30.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.31
  • blk.31.attn_norm.bias
    F32
    [8192]
  • blk.31.attn_norm.weight
    F32
    [8192]
  • blk.31.attn_norm_2.bias
    F32
    [8192]
  • blk.31.attn_norm_2.weight
    F32
    [8192]
  • blk.31.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.31.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.31.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.31.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.32
  • blk.32.attn_norm.bias
    F32
    [8192]
  • blk.32.attn_norm.weight
    F32
    [8192]
  • blk.32.attn_norm_2.bias
    F32
    [8192]
  • blk.32.attn_norm_2.weight
    F32
    [8192]
  • blk.32.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.32.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.32.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.32.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.33
  • blk.33.attn_norm.bias
    F32
    [8192]
  • blk.33.attn_norm.weight
    F32
    [8192]
  • blk.33.attn_norm_2.bias
    F32
    [8192]
  • blk.33.attn_norm_2.weight
    F32
    [8192]
  • blk.33.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.33.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.33.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.33.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.34
  • blk.34.attn_norm.bias
    F32
    [8192]
  • blk.34.attn_norm.weight
    F32
    [8192]
  • blk.34.attn_norm_2.bias
    F32
    [8192]
  • blk.34.attn_norm_2.weight
    F32
    [8192]
  • blk.34.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.34.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.34.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.34.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.35
  • blk.35.attn_norm.bias
    F32
    [8192]
  • blk.35.attn_norm.weight
    F32
    [8192]
  • blk.35.attn_norm_2.bias
    F32
    [8192]
  • blk.35.attn_norm_2.weight
    F32
    [8192]
  • blk.35.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.35.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.35.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.35.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.36
  • blk.36.attn_norm.bias
    F32
    [8192]
  • blk.36.attn_norm.weight
    F32
    [8192]
  • blk.36.attn_norm_2.bias
    F32
    [8192]
  • blk.36.attn_norm_2.weight
    F32
    [8192]
  • blk.36.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.36.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.36.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.36.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.37
  • blk.37.attn_norm.bias
    F32
    [8192]
  • blk.37.attn_norm.weight
    F32
    [8192]
  • blk.37.attn_norm_2.bias
    F32
    [8192]
  • blk.37.attn_norm_2.weight
    F32
    [8192]
  • blk.37.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.37.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.37.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.37.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.38
  • blk.38.attn_norm.bias
    F32
    [8192]
  • blk.38.attn_norm.weight
    F32
    [8192]
  • blk.38.attn_norm_2.bias
    F32
    [8192]
  • blk.38.attn_norm_2.weight
    F32
    [8192]
  • blk.38.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.38.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.38.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.38.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.39
  • blk.39.attn_norm.bias
    F32
    [8192]
  • blk.39.attn_norm.weight
    F32
    [8192]
  • blk.39.attn_norm_2.bias
    F32
    [8192]
  • blk.39.attn_norm_2.weight
    F32
    [8192]
  • blk.39.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.39.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.39.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.39.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.40
  • blk.40.attn_norm.bias
    F32
    [8192]
  • blk.40.attn_norm.weight
    F32
    [8192]
  • blk.40.attn_norm_2.bias
    F32
    [8192]
  • blk.40.attn_norm_2.weight
    F32
    [8192]
  • blk.40.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.40.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.40.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.40.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.41
  • blk.41.attn_norm.bias
    F32
    [8192]
  • blk.41.attn_norm.weight
    F32
    [8192]
  • blk.41.attn_norm_2.bias
    F32
    [8192]
  • blk.41.attn_norm_2.weight
    F32
    [8192]
  • blk.41.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.41.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.41.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.41.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.42
  • blk.42.attn_norm.bias
    F32
    [8192]
  • blk.42.attn_norm.weight
    F32
    [8192]
  • blk.42.attn_norm_2.bias
    F32
    [8192]
  • blk.42.attn_norm_2.weight
    F32
    [8192]
  • blk.42.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.42.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.42.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.42.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.43
  • blk.43.attn_norm.bias
    F32
    [8192]
  • blk.43.attn_norm.weight
    F32
    [8192]
  • blk.43.attn_norm_2.bias
    F32
    [8192]
  • blk.43.attn_norm_2.weight
    F32
    [8192]
  • blk.43.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.43.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.43.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.43.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.44
  • blk.44.attn_norm.bias
    F32
    [8192]
  • blk.44.attn_norm.weight
    F32
    [8192]
  • blk.44.attn_norm_2.bias
    F32
    [8192]
  • blk.44.attn_norm_2.weight
    F32
    [8192]
  • blk.44.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.44.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.44.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.44.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.45
  • blk.45.attn_norm.bias
    F32
    [8192]
  • blk.45.attn_norm.weight
    F32
    [8192]
  • blk.45.attn_norm_2.bias
    F32
    [8192]
  • blk.45.attn_norm_2.weight
    F32
    [8192]
  • blk.45.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.45.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.45.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.45.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.46
  • blk.46.attn_norm.bias
    F32
    [8192]
  • blk.46.attn_norm.weight
    F32
    [8192]
  • blk.46.attn_norm_2.bias
    F32
    [8192]
  • blk.46.attn_norm_2.weight
    F32
    [8192]
  • blk.46.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.46.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.46.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.46.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.47
  • blk.47.attn_norm.bias
    F32
    [8192]
  • blk.47.attn_norm.weight
    F32
    [8192]
  • blk.47.attn_norm_2.bias
    F32
    [8192]
  • blk.47.attn_norm_2.weight
    F32
    [8192]
  • blk.47.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.47.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.47.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.47.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.48
  • blk.48.attn_norm.bias
    F32
    [8192]
  • blk.48.attn_norm.weight
    F32
    [8192]
  • blk.48.attn_norm_2.bias
    F32
    [8192]
  • blk.48.attn_norm_2.weight
    F32
    [8192]
  • blk.48.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.48.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.48.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.48.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.49
  • blk.49.attn_norm.bias
    F32
    [8192]
  • blk.49.attn_norm.weight
    F32
    [8192]
  • blk.49.attn_norm_2.bias
    F32
    [8192]
  • blk.49.attn_norm_2.weight
    F32
    [8192]
  • blk.49.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.49.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.49.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.49.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.50
  • blk.50.attn_norm.bias
    F32
    [8192]
  • blk.50.attn_norm.weight
    F32
    [8192]
  • blk.50.attn_norm_2.bias
    F32
    [8192]
  • blk.50.attn_norm_2.weight
    F32
    [8192]
  • blk.50.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.50.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.50.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.50.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.51
  • blk.51.attn_norm.bias
    F32
    [8192]
  • blk.51.attn_norm.weight
    F32
    [8192]
  • blk.51.attn_norm_2.bias
    F32
    [8192]
  • blk.51.attn_norm_2.weight
    F32
    [8192]
  • blk.51.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.51.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.51.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.51.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.52
  • blk.52.attn_norm.bias
    F32
    [8192]
  • blk.52.attn_norm.weight
    F32
    [8192]
  • blk.52.attn_norm_2.bias
    F32
    [8192]
  • blk.52.attn_norm_2.weight
    F32
    [8192]
  • blk.52.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.52.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.52.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.52.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.53
  • blk.53.attn_norm.bias
    F32
    [8192]
  • blk.53.attn_norm.weight
    F32
    [8192]
  • blk.53.attn_norm_2.bias
    F32
    [8192]
  • blk.53.attn_norm_2.weight
    F32
    [8192]
  • blk.53.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.53.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.53.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.53.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.54
  • blk.54.attn_norm.bias
    F32
    [8192]
  • blk.54.attn_norm.weight
    F32
    [8192]
  • blk.54.attn_norm_2.bias
    F32
    [8192]
  • blk.54.attn_norm_2.weight
    F32
    [8192]
  • blk.54.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.54.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.54.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.54.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.55
  • blk.55.attn_norm.bias
    F32
    [8192]
  • blk.55.attn_norm.weight
    F32
    [8192]
  • blk.55.attn_norm_2.bias
    F32
    [8192]
  • blk.55.attn_norm_2.weight
    F32
    [8192]
  • blk.55.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.55.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.55.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.55.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.56
  • blk.56.attn_norm.bias
    F32
    [8192]
  • blk.56.attn_norm.weight
    F32
    [8192]
  • blk.56.attn_norm_2.bias
    F32
    [8192]
  • blk.56.attn_norm_2.weight
    F32
    [8192]
  • blk.56.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.56.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.56.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.56.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.57
  • blk.57.attn_norm.bias
    F32
    [8192]
  • blk.57.attn_norm.weight
    F32
    [8192]
  • blk.57.attn_norm_2.bias
    F32
    [8192]
  • blk.57.attn_norm_2.weight
    F32
    [8192]
  • blk.57.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.57.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.57.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.57.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.58
  • blk.58.attn_norm.bias
    F32
    [8192]
  • blk.58.attn_norm.weight
    F32
    [8192]
  • blk.58.attn_norm_2.bias
    F32
    [8192]
  • blk.58.attn_norm_2.weight
    F32
    [8192]
  • blk.58.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.58.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.58.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.58.ffn_up.weight
    Q4_0
    [8192, 32768]
    • blk.59
  • blk.59.attn_norm.bias
    F32
    [8192]
  • blk.59.attn_norm.weight
    F32
    [8192]
  • blk.59.attn_norm_2.bias
    F32
    [8192]
  • blk.59.attn_norm_2.weight
    F32
    [8192]
  • blk.59.attn_output.weight
    Q4_0
    [8192, 8192]
  • blk.59.attn_qkv.weight
    Q4_0
    [8192, 9216]
  • blk.59.ffn_down.weight
    Q4_0
    [32768, 8192]
  • blk.59.ffn_up.weight
    Q4_0
    [8192, 32768]
  • output.weight
    Q8_0
    [8192, 65024]
  • output_norm.bias
    F32
    [8192]
  • output_norm.weight
    F32
    [8192]