latest
66MB
Text embedding model (small) for input of size up to 8192 tokens
Embedding
716 Pulls Updated 4 months ago
465b654b2a5a · 66MB
-
general.architecturejina-bert-v2
-
general.file_typeF16
-
jina-bert-v2.attention.causalfalse
-
jina-bert-v2.attention.head_count8
-
jina-bert-v2.attention.layer_norm_epsilon1e-12
-
jina-bert-v2.block_count4
-
jina-bert-v2.context_length8192
-
jina-bert-v2.embedding_length512
-
jina-bert-v2.feed_forward_length2048
-
jina-bert-v2.pooling_typeCLS token pooling
-
tokenizer.ggml.add_bos_tokentrue
-
tokenizer.ggml.add_eos_tokentrue
-
tokenizer.ggml.cls_token_id101
-
tokenizer.ggml.mask_token_id103
-
tokenizer.ggml.modelbert
-
tokenizer.ggml.padding_token_id0
-
tokenizer.ggml.prejina-v2-en
-
tokenizer.ggml.seperator_token_id102
-
tokenizer.ggml.token_type[3 1 1 1 1 ...]
-
tokenizer.ggml.token_type_count2
-
tokenizer.ggml.tokens[[PAD] [unused0] [unused1] [unused2] [unused3] ...]
-
tokenizer.ggml.unknown_token_id100
-
NameTypeShape
-
token_embd_norm.biasF32[512]
-
token_embd_norm.weightF32[512]
-
token_embd.weightF16[512 30528]
-
blk.0.attn_output_norm.biasF32[512]
-
blk.0.attn_output_norm.weightF32[512]
-
blk.0.attn_output.biasF32[512]
-
blk.0.attn_output.weightF16[512 512]
-
blk.0.attn_k.biasF32[512]
-
blk.0.attn_k.weightF16[512 512]
-
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_gate.weightF16[512 2048]
-
blk.0.ffn_up.weightF16[512 2048]
-
blk.0.layer_output_norm.biasF32[512]
-
blk.0.layer_output_norm.weightF32[512]
-
blk.0.ffn_down.biasF32[512]
-
blk.0.ffn_down.weightF16[2048 512]
-
blk.1.attn_output_norm.biasF32[512]
-
blk.1.attn_output_norm.weightF32[512]
-
blk.1.attn_output.biasF32[512]
-
blk.1.attn_output.weightF16[512 512]
-
blk.1.attn_k.biasF32[512]
-
blk.1.attn_k.weightF16[512 512]
-
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_gate.weightF16[512 2048]
-
blk.1.ffn_up.weightF16[512 2048]
-
blk.1.layer_output_norm.biasF32[512]
-
blk.1.layer_output_norm.weightF32[512]
-
blk.1.ffn_down.biasF32[512]
-
blk.1.ffn_down.weightF16[2048 512]
-
blk.2.attn_output_norm.biasF32[512]
-
blk.2.attn_output_norm.weightF32[512]
-
blk.2.attn_output.biasF32[512]
-
blk.2.attn_output.weightF16[512 512]
-
blk.2.attn_k.biasF32[512]
-
blk.2.attn_k.weightF16[512 512]
-
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_gate.weightF16[512 2048]
-
blk.2.ffn_up.weightF16[512 2048]
-
blk.2.layer_output_norm.biasF32[512]
-
blk.2.layer_output_norm.weightF32[512]
-
blk.2.ffn_down.biasF32[512]
-
blk.2.ffn_down.weightF16[2048 512]
-
blk.3.attn_output_norm.biasF32[512]
-
blk.3.attn_output_norm.weightF32[512]
-
blk.3.attn_output.biasF32[512]
-
blk.3.attn_output.weightF16[512 512]
-
blk.3.attn_k.biasF32[512]
-
blk.3.attn_k.weightF16[512 512]
-
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_gate.weightF16[512 2048]
-
blk.3.ffn_up.weightF16[512 2048]
-
blk.3.layer_output_norm.biasF32[512]
-
blk.3.layer_output_norm.weightF32[512]
-
blk.3.ffn_down.biasF32[512]
-
blk.3.ffn_down.weightF16[2048 512]
-
token_types.weightF32[512 2]
Metadata
Tensors
blk.0
blk.1
blk.2
blk.3