Sentence-transformers model that can be used for tasks like clustering or semantic search.
embedding
278m
6,463 Pulls Updated 3 months ago
aa99ebfc77f4 · 563MB
-
bert.attention.causalfalse
-
bert.attention.head_count12
-
bert.attention.layer_norm_epsilon1e-05
-
bert.block_count12
-
bert.context_length512
-
bert.embedding_length768
-
bert.feed_forward_length3072
-
bert.pooling_type1
-
general.architecturebert
-
general.basenameparaphrase-multilingual-mpnet-base
-
general.file_type1
-
general.languages[multilingual, ar, bg, ca, cs, ...]
-
general.licenseapache-2.0
-
general.nameParaphrase Multilingual Mpnet Base v2
-
general.quantization_version2
-
general.size_label277M
-
general.tags[sentence-transformers, feature-extraction, sentence-similarity, transformers, sentence-similarity]
-
general.typemodel
-
general.versionv2
-
tokenizer.ggml.add_bos_tokentrue
-
tokenizer.ggml.add_eos_tokentrue
-
tokenizer.ggml.add_space_prefixtrue
-
tokenizer.ggml.bos_token_id0
-
tokenizer.ggml.cls_token_id0
-
tokenizer.ggml.eos_token_id2
-
tokenizer.ggml.mask_token_id250001
-
tokenizer.ggml.modelt5
-
tokenizer.ggml.padding_token_id1
-
tokenizer.ggml.predefault
-
tokenizer.ggml.precompiled_charsmap[0, 180, 2, 0, 0, ...]
-
tokenizer.ggml.remove_extra_whitespacestrue
-
tokenizer.ggml.scores[0, 0, 0, 0, -3.4635427, ...]
-
tokenizer.ggml.seperator_token_id2
-
tokenizer.ggml.token_type[3, 3, 3, 2, 1, ...]
-
tokenizer.ggml.token_type_count1
-
tokenizer.ggml.tokens[<s>, <pad>, </s>, <unk>, ,, ...]
-
tokenizer.ggml.unknown_token_id3
-
NameTypeShape
-
token_embd.weightF16[768, 250002]
-
blk.0.attn_k.biasF32[768]
-
blk.0.attn_k.weightF16[768, 768]
-
blk.0.attn_output.biasF32[768]
-
blk.0.attn_output.weightF16[768, 768]
-
blk.0.attn_output_norm.biasF32[768]
-
blk.0.attn_output_norm.weightF32[768]
-
blk.0.attn_q.biasF32[768]
-
blk.0.attn_q.weightF16[768, 768]
-
blk.0.attn_v.biasF32[768]
-
blk.0.attn_v.weightF16[768, 768]
-
blk.0.ffn_down.biasF32[768]
-
blk.0.ffn_down.weightF16[3072, 768]
-
blk.0.ffn_up.biasF32[3072]
-
blk.0.ffn_up.weightF16[768, 3072]
-
blk.0.layer_output_norm.biasF32[768]
-
blk.0.layer_output_norm.weightF32[768]
-
blk.1.attn_k.biasF32[768]
-
blk.1.attn_k.weightF16[768, 768]
-
blk.1.attn_output.biasF32[768]
-
blk.1.attn_output.weightF16[768, 768]
-
blk.1.attn_output_norm.biasF32[768]
-
blk.1.attn_output_norm.weightF32[768]
-
blk.1.attn_q.biasF32[768]
-
blk.1.attn_q.weightF16[768, 768]
-
blk.1.attn_v.biasF32[768]
-
blk.1.attn_v.weightF16[768, 768]
-
blk.1.ffn_down.biasF32[768]
-
blk.1.ffn_down.weightF16[3072, 768]
-
blk.1.ffn_up.biasF32[3072]
-
blk.1.ffn_up.weightF16[768, 3072]
-
blk.1.layer_output_norm.biasF32[768]
-
blk.1.layer_output_norm.weightF32[768]
-
blk.2.attn_k.biasF32[768]
-
blk.2.attn_k.weightF16[768, 768]
-
blk.2.attn_output.biasF32[768]
-
blk.2.attn_output.weightF16[768, 768]
-
blk.2.attn_output_norm.biasF32[768]
-
blk.2.attn_output_norm.weightF32[768]
-
blk.2.attn_q.biasF32[768]
-
blk.2.attn_q.weightF16[768, 768]
-
blk.2.attn_v.biasF32[768]
-
blk.2.attn_v.weightF16[768, 768]
-
blk.2.ffn_down.biasF32[768]
-
blk.2.ffn_down.weightF16[3072, 768]
-
blk.2.ffn_up.biasF32[3072]
-
blk.2.ffn_up.weightF16[768, 3072]
-
blk.2.layer_output_norm.biasF32[768]
-
blk.2.layer_output_norm.weightF32[768]
-
blk.3.attn_k.biasF32[768]
-
blk.3.attn_k.weightF16[768, 768]
-
blk.3.attn_output.biasF32[768]
-
blk.3.attn_output.weightF16[768, 768]
-
blk.3.attn_output_norm.biasF32[768]
-
blk.3.attn_output_norm.weightF32[768]
-
blk.3.attn_q.biasF32[768]
-
blk.3.attn_q.weightF16[768, 768]
-
blk.3.attn_v.biasF32[768]
-
blk.3.attn_v.weightF16[768, 768]
-
blk.3.ffn_down.biasF32[768]
-
blk.3.ffn_down.weightF16[3072, 768]
-
blk.3.ffn_up.biasF32[3072]
-
blk.3.ffn_up.weightF16[768, 3072]
-
blk.3.layer_output_norm.biasF32[768]
-
blk.3.layer_output_norm.weightF32[768]
-
blk.4.attn_k.biasF32[768]
-
blk.4.attn_k.weightF16[768, 768]
-
blk.4.attn_output.biasF32[768]
-
blk.4.attn_output.weightF16[768, 768]
-
blk.4.attn_output_norm.biasF32[768]
-
blk.4.attn_output_norm.weightF32[768]
-
blk.4.attn_q.biasF32[768]
-
blk.4.attn_q.weightF16[768, 768]
-
blk.4.attn_v.biasF32[768]
-
blk.4.attn_v.weightF16[768, 768]
-
blk.4.ffn_down.biasF32[768]
-
blk.4.ffn_down.weightF16[3072, 768]
-
blk.4.ffn_up.biasF32[3072]
-
blk.4.ffn_up.weightF16[768, 3072]
-
blk.4.layer_output_norm.biasF32[768]
-
blk.4.layer_output_norm.weightF32[768]
-
blk.5.attn_k.biasF32[768]
-
blk.5.attn_k.weightF16[768, 768]
-
blk.5.attn_output.biasF32[768]
-
blk.5.attn_output.weightF16[768, 768]
-
blk.5.attn_output_norm.biasF32[768]
-
blk.5.attn_output_norm.weightF32[768]
-
blk.5.attn_q.biasF32[768]
-
blk.5.attn_q.weightF16[768, 768]
-
blk.5.attn_v.biasF32[768]
-
blk.5.attn_v.weightF16[768, 768]
-
blk.5.ffn_down.biasF32[768]
-
blk.5.ffn_down.weightF16[3072, 768]
-
blk.5.ffn_up.biasF32[3072]
-
blk.5.ffn_up.weightF16[768, 3072]
-
blk.5.layer_output_norm.biasF32[768]
-
blk.5.layer_output_norm.weightF32[768]
-
blk.6.attn_k.biasF32[768]
-
blk.6.attn_k.weightF16[768, 768]
-
blk.6.attn_output.biasF32[768]
-
blk.6.attn_output.weightF16[768, 768]
-
blk.6.attn_output_norm.biasF32[768]
-
blk.6.attn_output_norm.weightF32[768]
-
blk.6.attn_q.biasF32[768]
-
blk.6.attn_q.weightF16[768, 768]
-
blk.6.attn_v.biasF32[768]
-
blk.6.attn_v.weightF16[768, 768]
-
blk.6.ffn_down.biasF32[768]
-
blk.6.ffn_down.weightF16[3072, 768]
-
blk.6.ffn_up.biasF32[3072]
-
blk.6.ffn_up.weightF16[768, 3072]
-
blk.6.layer_output_norm.biasF32[768]
-
blk.6.layer_output_norm.weightF32[768]
-
blk.7.attn_k.biasF32[768]
-
blk.7.attn_k.weightF16[768, 768]
-
blk.7.attn_output.biasF32[768]
-
blk.7.attn_output.weightF16[768, 768]
-
blk.7.attn_output_norm.biasF32[768]
-
blk.7.attn_output_norm.weightF32[768]
-
blk.7.attn_q.biasF32[768]
-
blk.7.attn_q.weightF16[768, 768]
-
blk.7.attn_v.biasF32[768]
-
blk.7.attn_v.weightF16[768, 768]
-
blk.7.ffn_down.biasF32[768]
-
blk.7.ffn_down.weightF16[3072, 768]
-
blk.7.ffn_up.biasF32[3072]
-
blk.7.ffn_up.weightF16[768, 3072]
-
blk.7.layer_output_norm.biasF32[768]
-
blk.7.layer_output_norm.weightF32[768]
-
blk.8.attn_k.biasF32[768]
-
blk.8.attn_k.weightF16[768, 768]
-
blk.8.attn_output.biasF32[768]
-
blk.8.attn_output.weightF16[768, 768]
-
blk.8.attn_output_norm.biasF32[768]
-
blk.8.attn_output_norm.weightF32[768]
-
blk.8.attn_q.biasF32[768]
-
blk.8.attn_q.weightF16[768, 768]
-
blk.8.attn_v.biasF32[768]
-
blk.8.attn_v.weightF16[768, 768]
-
blk.8.ffn_down.biasF32[768]
-
blk.8.ffn_down.weightF16[3072, 768]
-
blk.8.ffn_up.biasF32[3072]
-
blk.8.ffn_up.weightF16[768, 3072]
-
blk.8.layer_output_norm.biasF32[768]
-
blk.8.layer_output_norm.weightF32[768]
-
blk.9.attn_k.biasF32[768]
-
blk.9.attn_k.weightF16[768, 768]
-
blk.9.attn_output.biasF32[768]
-
blk.9.attn_output.weightF16[768, 768]
-
blk.9.attn_output_norm.biasF32[768]
-
blk.9.attn_output_norm.weightF32[768]
-
blk.9.attn_q.biasF32[768]
-
blk.9.attn_q.weightF16[768, 768]
-
blk.9.attn_v.biasF32[768]
-
blk.9.attn_v.weightF16[768, 768]
-
blk.9.ffn_down.biasF32[768]
-
blk.9.ffn_down.weightF16[3072, 768]
-
blk.9.ffn_up.biasF32[3072]
-
blk.9.ffn_up.weightF16[768, 3072]
-
blk.9.layer_output_norm.biasF32[768]
-
blk.9.layer_output_norm.weightF32[768]
-
blk.10.attn_k.biasF32[768]
-
blk.10.attn_k.weightF16[768, 768]
-
blk.10.attn_output.biasF32[768]
-
blk.10.attn_output.weightF16[768, 768]
-
blk.10.attn_output_norm.biasF32[768]
-
blk.10.attn_output_norm.weightF32[768]
-
blk.10.attn_q.biasF32[768]
-
blk.10.attn_q.weightF16[768, 768]
-
blk.10.attn_v.biasF32[768]
-
blk.10.attn_v.weightF16[768, 768]
-
blk.10.ffn_down.biasF32[768]
-
blk.10.ffn_down.weightF16[3072, 768]
-
blk.10.ffn_up.biasF32[3072]
-
blk.10.ffn_up.weightF16[768, 3072]
-
blk.10.layer_output_norm.biasF32[768]
-
blk.10.layer_output_norm.weightF32[768]
-
blk.11.attn_k.biasF32[768]
-
blk.11.attn_k.weightF16[768, 768]
-
blk.11.attn_output.biasF32[768]
-
blk.11.attn_output.weightF16[768, 768]
-
blk.11.attn_output_norm.biasF32[768]
-
blk.11.attn_output_norm.weightF32[768]
-
blk.11.attn_q.biasF32[768]
-
blk.11.attn_q.weightF16[768, 768]
-
blk.11.attn_v.biasF32[768]
-
blk.11.attn_v.weightF16[768, 768]
-
blk.11.ffn_down.biasF32[768]
-
blk.11.ffn_down.weightF16[3072, 768]
-
blk.11.ffn_up.biasF32[3072]
-
blk.11.ffn_up.weightF16[768, 3072]
-
blk.11.layer_output_norm.biasF32[768]
-
blk.11.layer_output_norm.weightF32[768]
-
position_embd.weightF32[768, 512]
-
token_embd_norm.biasF32[768]
-
token_embd_norm.weightF32[768]
-
token_types.weightF32[768]
Metadata
Tensor
blk.0
blk.1
blk.2
blk.3
blk.4
blk.5
blk.6
blk.7
blk.8
blk.9
blk.10
blk.11