Moka-AI Massive Mixed Embedding
embedding
3,096 Pulls Updated 8 months ago
ed780aa4e802 · 48MB
-
bert.attention.causalfalse
-
bert.attention.head_count8
-
bert.attention.layer_norm_epsilon1e-12
-
bert.block_count4
-
bert.context_length512
-
bert.embedding_length512
-
bert.feed_forward_length2048
-
bert.pooling_type1
-
general.architecturebert
-
general.file_type1
-
general.namem3e-small
-
tokenizer.ggml.bos_token_id101
-
tokenizer.ggml.cls_token_id101
-
tokenizer.ggml.eos_token_id102
-
tokenizer.ggml.mask_token_id103
-
tokenizer.ggml.modelbert
-
tokenizer.ggml.padding_token_id0
-
tokenizer.ggml.scores[-1000, -1000, -1000, -1000, -1000, ...]
-
tokenizer.ggml.seperator_token_id102
-
tokenizer.ggml.token_type[3, 1, 1, 1, 1, ...]
-
tokenizer.ggml.token_type_count2
-
tokenizer.ggml.tokens[[PAD], [unused1], [unused2], [unused3], [unused4], ...]
-
tokenizer.ggml.unknown_token_id100
-
NameTypeShape
-
token_embd.weightF16[512, 21128]
-
blk.0.attn_k.biasF32[512]
-
blk.0.attn_k.weightF16[512, 512]
-
blk.0.attn_output.biasF32[512]
-
blk.0.attn_output.weightF16[512, 512]
-
blk.0.attn_output_norm.biasF32[512]
-
blk.0.attn_output_norm.weightF32[512]
-
blk.0.attn_q.biasF32[512]
-
blk.0.attn_q.weightF16[512, 512]
-
blk.0.attn_v.biasF32[512]
-
blk.0.attn_v.weightF16[512, 512]
-
blk.0.ffn_down.biasF32[512]
-
blk.0.ffn_down.weightF16[2048, 512]
-
blk.0.ffn_up.biasF32[2048]
-
blk.0.ffn_up.weightF16[512, 2048]
-
blk.0.layer_output_norm.biasF32[512]
-
blk.0.layer_output_norm.weightF32[512]
-
blk.1.attn_k.biasF32[512]
-
blk.1.attn_k.weightF16[512, 512]
-
blk.1.attn_output.biasF32[512]
-
blk.1.attn_output.weightF16[512, 512]
-
blk.1.attn_output_norm.biasF32[512]
-
blk.1.attn_output_norm.weightF32[512]
-
blk.1.attn_q.biasF32[512]
-
blk.1.attn_q.weightF16[512, 512]
-
blk.1.attn_v.biasF32[512]
-
blk.1.attn_v.weightF16[512, 512]
-
blk.1.ffn_down.biasF32[512]
-
blk.1.ffn_down.weightF16[2048, 512]
-
blk.1.ffn_up.biasF32[2048]
-
blk.1.ffn_up.weightF16[512, 2048]
-
blk.1.layer_output_norm.biasF32[512]
-
blk.1.layer_output_norm.weightF32[512]
-
blk.2.attn_k.biasF32[512]
-
blk.2.attn_k.weightF16[512, 512]
-
blk.2.attn_output.biasF32[512]
-
blk.2.attn_output.weightF16[512, 512]
-
blk.2.attn_output_norm.biasF32[512]
-
blk.2.attn_output_norm.weightF32[512]
-
blk.2.attn_q.biasF32[512]
-
blk.2.attn_q.weightF16[512, 512]
-
blk.2.attn_v.biasF32[512]
-
blk.2.attn_v.weightF16[512, 512]
-
blk.2.ffn_down.biasF32[512]
-
blk.2.ffn_down.weightF16[2048, 512]
-
blk.2.ffn_up.biasF32[2048]
-
blk.2.ffn_up.weightF16[512, 2048]
-
blk.2.layer_output_norm.biasF32[512]
-
blk.2.layer_output_norm.weightF32[512]
-
blk.3.attn_k.biasF32[512]
-
blk.3.attn_k.weightF16[512, 512]
-
blk.3.attn_output.biasF32[512]
-
blk.3.attn_output.weightF16[512, 512]
-
blk.3.attn_output_norm.biasF32[512]
-
blk.3.attn_output_norm.weightF32[512]
-
blk.3.attn_q.biasF32[512]
-
blk.3.attn_q.weightF16[512, 512]
-
blk.3.attn_v.biasF32[512]
-
blk.3.attn_v.weightF16[512, 512]
-
blk.3.ffn_down.biasF32[512]
-
blk.3.ffn_down.weightF16[2048, 512]
-
blk.3.ffn_up.biasF32[2048]
-
blk.3.ffn_up.weightF16[512, 2048]
-
blk.3.layer_output_norm.biasF32[512]
-
blk.3.layer_output_norm.weightF32[512]
-
position_embd.weightF16[512, 512]
-
token_embd_norm.biasF32[512]
-
token_embd_norm.weightF32[512]
-
token_types.weightF32[512, 2]
Metadata
Tensor
blk.0
blk.1
blk.2
blk.3