sppandita85/ dubai-tourism-llm:latest

2 1 month ago

trained on 22k tokens about dubai tourism

... /
model
be0cab0936c9 · 76MB
    Metadata
  • general.architecture
    gpt2
  • general.file_type
    (!unknown_file_type 1024!)
  • gpt2.attention.head_count
    8
  • gpt2.attention.layer_norm_epsilon
    1e-05
  • gpt2.block_count
    4
  • gpt2.context_length
    128
  • gpt2.embedding_length
    256
  • gpt2.feed_forward_length
    1024
  • tokenizer.ggml.bos_token_id
    101
  • tokenizer.ggml.eos_token_id
    102
  • tokenizer.ggml.model
    bert
  • tokenizer.ggml.padding_token_id
    0
  • tokenizer.ggml.token_type
    [3, 1, 1, 1, 1, ...]
  • tokenizer.ggml.tokens
    [[PAD], [unused0], [unused1], [unused2], [unused3], ...]
  • tokenizer.ggml.unknown_token_id
    100
    • Tensor
  • token_embd.weight
    F32
    [256, 30522]
    • blk.0
  • blk.0.attn_norm.bias
    F32
    [256]
  • blk.0.attn_norm.weight
    F32
    [256]
  • blk.0.attn_output.bias
    F32
    [256]
  • blk.0.attn_output.weight
    F32
    [256, 256]
  • blk.0.attn_qkv.bias
    F32
    [768]
  • blk.0.attn_qkv.weight
    F32
    [256, 768]
  • blk.0.ffn_down.bias
    F32
    [256]
  • blk.0.ffn_down.weight
    F32
    [1024, 256]
  • blk.0.ffn_norm.bias
    F32
    [256]
  • blk.0.ffn_norm.weight
    F32
    [256]
  • blk.0.ffn_up.bias
    F32
    [1024]
  • blk.0.ffn_up.weight
    F32
    [256, 1024]
    • blk.1
  • blk.1.attn_norm.bias
    F32
    [256]
  • blk.1.attn_norm.weight
    F32
    [256]
  • blk.1.attn_output.bias
    F32
    [256]
  • blk.1.attn_output.weight
    F32
    [256, 256]
  • blk.1.attn_qkv.bias
    F32
    [768]
  • blk.1.attn_qkv.weight
    F32
    [256, 768]
  • blk.1.ffn_down.bias
    F32
    [256]
  • blk.1.ffn_down.weight
    F32
    [1024, 256]
  • blk.1.ffn_norm.bias
    F32
    [256]
  • blk.1.ffn_norm.weight
    F32
    [256]
  • blk.1.ffn_up.bias
    F32
    [1024]
  • blk.1.ffn_up.weight
    F32
    [256, 1024]
    • blk.2
  • blk.2.attn_norm.bias
    F32
    [256]
  • blk.2.attn_norm.weight
    F32
    [256]
  • blk.2.attn_output.bias
    F32
    [256]
  • blk.2.attn_output.weight
    F32
    [256, 256]
  • blk.2.attn_qkv.bias
    F32
    [768]
  • blk.2.attn_qkv.weight
    F32
    [256, 768]
  • blk.2.ffn_down.bias
    F32
    [256]
  • blk.2.ffn_down.weight
    F32
    [1024, 256]
  • blk.2.ffn_norm.bias
    F32
    [256]
  • blk.2.ffn_norm.weight
    F32
    [256]
  • blk.2.ffn_up.bias
    F32
    [1024]
  • blk.2.ffn_up.weight
    F32
    [256, 1024]
    • blk.3
  • blk.3.attn_norm.bias
    F32
    [256]
  • blk.3.attn_norm.weight
    F32
    [256]
  • blk.3.attn_output.bias
    F32
    [256]
  • blk.3.attn_output.weight
    F32
    [256, 256]
  • blk.3.attn_qkv.bias
    F32
    [768]
  • blk.3.attn_qkv.weight
    F32
    [256, 768]
  • blk.3.ffn_down.bias
    F32
    [256]
  • blk.3.ffn_down.weight
    F32
    [1024, 256]
  • blk.3.ffn_norm.bias
    F32
    [256]
  • blk.3.ffn_norm.weight
    F32
    [256]
  • blk.3.ffn_up.bias
    F32
    [1024]
  • blk.3.ffn_up.weight
    F32
    [256, 1024]
  • output.weight
    F32
    [256, 30522]
  • output_norm.bias
    F32
    [256]
  • position_embd.weight
    F32
    [256, 128]
  • output_norm.weight
    F32
    [256]