treyleo16/ small-minded:260k

67 1 week ago

Like a little baby with a small brain.

0.0001m 0.00077m 0.001m 0.26m 1m
... /
model
d23eccb1c045 · 1.8MB
    Metadata
  • general.architecture
    llama
  • general.file_type
    F32
  • llama.attention.head_count
    2
  • llama.attention.head_count_kv
    2
  • llama.attention.layer_norm_rms_epsilon
    1e-05
  • llama.block_count
    2
  • llama.context_length
    512
  • llama.embedding_length
    8
  • llama.feed_forward_length
    72
  • llama.rope.dimension_count
    4
  • llama.rope.freq_base
    10000
  • llama.vocab_size
    32000
  • tokenizer.ggml.add_bos_token
    true
  • tokenizer.ggml.add_eos_token
    false
  • tokenizer.ggml.bos_token_id
    1
  • tokenizer.ggml.eos_token_id
    2
  • tokenizer.ggml.model
    llama
  • tokenizer.ggml.padding_token_id
    0
  • tokenizer.ggml.scores
    [0, 0, 0, 0, 0, ...]
  • tokenizer.ggml.token_type
    [2, 3, 3, 6, 6, ...]
  • tokenizer.ggml.tokens
    [<unk>, <s>, </s>, <0x00>, <0x01>, ...]
  • tokenizer.ggml.unknown_token_id
    0
    • Tensor
  • token_embd.weight
    F32
    [8, 32000]
    • blk.0
  • blk.0.attn_k.weight
    F32
    [8, 8]
  • blk.0.attn_norm.weight
    F32
    [8]
  • blk.0.attn_output.weight
    F32
    [8, 8]
  • blk.0.attn_q.weight
    F32
    [8, 8]
  • blk.0.attn_v.weight
    F32
    [8, 8]
  • blk.0.ffn_down.weight
    F32
    [72, 8]
  • blk.0.ffn_gate.weight
    F32
    [8, 72]
  • blk.0.ffn_norm.weight
    F32
    [8]
  • blk.0.ffn_up.weight
    F32
    [8, 72]
    • blk.1
  • blk.1.attn_k.weight
    F32
    [8, 8]
  • blk.1.attn_norm.weight
    F32
    [8]
  • blk.1.attn_output.weight
    F32
    [8, 8]
  • blk.1.attn_q.weight
    F32
    [8, 8]
  • blk.1.attn_v.weight
    F32
    [8, 8]
  • blk.1.ffn_down.weight
    F32
    [72, 8]
  • blk.1.ffn_gate.weight
    F32
    [8, 72]
  • blk.1.ffn_norm.weight
    F32
    [8]
  • blk.1.ffn_up.weight
    F32
    [8, 72]
  • output_norm.weight
    F32
    [8]