Language model for Russian
125m
356m
760m
178 Pulls Updated 5 months ago
9f60778bacfe · 904MB
-
general.architecturegpt2gpt2
-
general.file_type77
-
general.nameRugpt3Large_Based_On_Gpt2Rugpt3Large_Based_On_Gpt2
-
general.organizationAi ForeverAi Forever
-
general.quantization_version22
-
general.size_label837M837M
-
general.typemodelmodel
-
gpt2.attention.head_count1616
-
gpt2.attention.layer_norm_epsilon1e-051e-05
-
gpt2.block_count2424
-
gpt2.context_length20482048
-
gpt2.embedding_length15361536
-
gpt2.feed_forward_length61446144
-
tokenizer.ggml.add_bos_tokenfalsefalse
-
tokenizer.ggml.bos_token_id11
-
tokenizer.ggml.eos_token_id22
-
tokenizer.ggml.mask_token_id44
-
tokenizer.ggml.merges[Ġ Ð, Ð ¾, Ð µ, Ð °, Ð ¸, ...][Ġ Ð, Ð ¾, Ð µ, Ð °, Ð ¸, ...]
-
tokenizer.ggml.modelgpt2gpt2
-
tokenizer.ggml.padding_token_id00
-
tokenizer.ggml.token_type[3, 3, 3, 3, 3, ...][3, 3, 3, 3, 3, ...]
-
tokenizer.ggml.tokens[<pad>, <s>, </s>, <unk>, <mask>, ...][<pad>, <s>, </s>, <unk>, <mask>, ...]
-
tokenizer.ggml.unknown_token_id33
-
token_embd.weightQ8_0[1536, 50257]
-
blk.0.attn_norm.biasF32[1536]
-
blk.0.attn_norm.weightF32[1536]
-
blk.0.attn_output.biasF32[1536]
-
blk.0.attn_output.weightQ8_0[1536, 1536]
-
blk.0.attn_qkv.biasF32[4608]
-
blk.0.attn_qkv.weightQ8_0[1536, 4608]
-
blk.0.ffn_down.biasF32[1536]
-
blk.0.ffn_down.weightQ8_0[6144, 1536]
-
blk.0.ffn_norm.biasF32[1536]
-
blk.0.ffn_norm.weightF32[1536]
-
blk.0.ffn_up.biasF32[6144]
-
blk.0.ffn_up.weightQ8_0[1536, 6144]
-
blk.1.attn_norm.biasF32[1536]
-
blk.1.attn_norm.weightF32[1536]
-
blk.1.attn_output.biasF32[1536]
-
blk.1.attn_output.weightQ8_0[1536, 1536]
-
blk.1.attn_qkv.biasF32[4608]
-
blk.1.attn_qkv.weightQ8_0[1536, 4608]
-
blk.1.ffn_down.biasF32[1536]
-
blk.1.ffn_down.weightQ8_0[6144, 1536]
-
blk.1.ffn_norm.biasF32[1536]
-
blk.1.ffn_norm.weightF32[1536]
-
blk.1.ffn_up.biasF32[6144]
-
blk.1.ffn_up.weightQ8_0[1536, 6144]
-
blk.2.attn_norm.biasF32[1536]
-
blk.2.attn_norm.weightF32[1536]
-
blk.2.attn_output.biasF32[1536]
-
blk.2.attn_output.weightQ8_0[1536, 1536]
-
blk.2.attn_qkv.biasF32[4608]
-
blk.2.attn_qkv.weightQ8_0[1536, 4608]
-
blk.2.ffn_down.biasF32[1536]
-
blk.2.ffn_down.weightQ8_0[6144, 1536]
-
blk.2.ffn_norm.biasF32[1536]
-
blk.2.ffn_norm.weightF32[1536]
-
blk.2.ffn_up.biasF32[6144]
-
blk.2.ffn_up.weightQ8_0[1536, 6144]
-
blk.3.attn_norm.biasF32[1536]
-
blk.3.attn_norm.weightF32[1536]
-
blk.3.attn_output.biasF32[1536]
-
blk.3.attn_output.weightQ8_0[1536, 1536]
-
blk.3.attn_qkv.biasF32[4608]
-
blk.3.attn_qkv.weightQ8_0[1536, 4608]
-
blk.3.ffn_down.biasF32[1536]
-
blk.3.ffn_down.weightQ8_0[6144, 1536]
-
blk.3.ffn_norm.biasF32[1536]
-
blk.3.ffn_norm.weightF32[1536]
-
blk.3.ffn_up.biasF32[6144]
-
blk.3.ffn_up.weightQ8_0[1536, 6144]
-
blk.4.attn_norm.biasF32[1536]
-
blk.4.attn_norm.weightF32[1536]
-
blk.4.attn_output.biasF32[1536]
-
blk.4.attn_output.weightQ8_0[1536, 1536]
-
blk.4.attn_qkv.biasF32[4608]
-
blk.4.attn_qkv.weightQ8_0[1536, 4608]
-
blk.4.ffn_down.biasF32[1536]
-
blk.4.ffn_down.weightQ8_0[6144, 1536]
-
blk.4.ffn_norm.biasF32[1536]
-
blk.4.ffn_norm.weightF32[1536]
-
blk.4.ffn_up.biasF32[6144]
-
blk.4.ffn_up.weightQ8_0[1536, 6144]
-
blk.5.attn_norm.biasF32[1536]
-
blk.5.attn_norm.weightF32[1536]
-
blk.5.attn_output.biasF32[1536]
-
blk.5.attn_output.weightQ8_0[1536, 1536]
-
blk.5.attn_qkv.biasF32[4608]
-
blk.5.attn_qkv.weightQ8_0[1536, 4608]
-
blk.5.ffn_down.biasF32[1536]
-
blk.5.ffn_down.weightQ8_0[6144, 1536]
-
blk.5.ffn_norm.biasF32[1536]
-
blk.5.ffn_norm.weightF32[1536]
-
blk.5.ffn_up.biasF32[6144]
-
blk.5.ffn_up.weightQ8_0[1536, 6144]
-
blk.6.attn_norm.biasF32[1536]
-
blk.6.attn_norm.weightF32[1536]
-
blk.6.attn_output.biasF32[1536]
-
blk.6.attn_output.weightQ8_0[1536, 1536]
-
blk.6.attn_qkv.biasF32[4608]
-
blk.6.attn_qkv.weightQ8_0[1536, 4608]
-
blk.6.ffn_down.biasF32[1536]
-
blk.6.ffn_down.weightQ8_0[6144, 1536]
-
blk.6.ffn_norm.biasF32[1536]
-
blk.6.ffn_norm.weightF32[1536]
-
blk.6.ffn_up.biasF32[6144]
-
blk.6.ffn_up.weightQ8_0[1536, 6144]
-
blk.7.attn_norm.biasF32[1536]
-
blk.7.attn_norm.weightF32[1536]
-
blk.7.attn_output.biasF32[1536]
-
blk.7.attn_output.weightQ8_0[1536, 1536]
-
blk.7.attn_qkv.biasF32[4608]
-
blk.7.attn_qkv.weightQ8_0[1536, 4608]
-
blk.7.ffn_down.biasF32[1536]
-
blk.7.ffn_down.weightQ8_0[6144, 1536]
-
blk.7.ffn_norm.biasF32[1536]
-
blk.7.ffn_norm.weightF32[1536]
-
blk.7.ffn_up.biasF32[6144]
-
blk.7.ffn_up.weightQ8_0[1536, 6144]
-
blk.8.attn_norm.biasF32[1536]
-
blk.8.attn_norm.weightF32[1536]
-
blk.8.attn_output.biasF32[1536]
-
blk.8.attn_output.weightQ8_0[1536, 1536]
-
blk.8.attn_qkv.biasF32[4608]
-
blk.8.attn_qkv.weightQ8_0[1536, 4608]
-
blk.8.ffn_down.biasF32[1536]
-
blk.8.ffn_down.weightQ8_0[6144, 1536]
-
blk.8.ffn_norm.biasF32[1536]
-
blk.8.ffn_norm.weightF32[1536]
-
blk.8.ffn_up.biasF32[6144]
-
blk.8.ffn_up.weightQ8_0[1536, 6144]
-
blk.9.attn_norm.biasF32[1536]
-
blk.9.attn_norm.weightF32[1536]
-
blk.9.attn_output.biasF32[1536]
-
blk.9.attn_output.weightQ8_0[1536, 1536]
-
blk.9.attn_qkv.biasF32[4608]
-
blk.9.attn_qkv.weightQ8_0[1536, 4608]
-
blk.9.ffn_down.biasF32[1536]
-
blk.9.ffn_down.weightQ8_0[6144, 1536]
-
blk.9.ffn_norm.biasF32[1536]
-
blk.9.ffn_norm.weightF32[1536]
-
blk.9.ffn_up.biasF32[6144]
-
blk.9.ffn_up.weightQ8_0[1536, 6144]
-
blk.10.attn_norm.biasF32[1536]
-
blk.10.attn_norm.weightF32[1536]
-
blk.10.attn_output.biasF32[1536]
-
blk.10.attn_output.weightQ8_0[1536, 1536]
-
blk.10.attn_qkv.biasF32[4608]
-
blk.10.attn_qkv.weightQ8_0[1536, 4608]
-
blk.10.ffn_down.biasF32[1536]
-
blk.10.ffn_down.weightQ8_0[6144, 1536]
-
blk.10.ffn_norm.biasF32[1536]
-
blk.10.ffn_norm.weightF32[1536]
-
blk.10.ffn_up.biasF32[6144]
-
blk.10.ffn_up.weightQ8_0[1536, 6144]
-
blk.11.attn_norm.biasF32[1536]
-
blk.11.attn_norm.weightF32[1536]
-
blk.11.attn_output.biasF32[1536]
-
blk.11.attn_output.weightQ8_0[1536, 1536]
-
blk.11.attn_qkv.biasF32[4608]
-
blk.11.attn_qkv.weightQ8_0[1536, 4608]
-
blk.11.ffn_down.biasF32[1536]
-
blk.11.ffn_down.weightQ8_0[6144, 1536]
-
blk.11.ffn_norm.biasF32[1536]
-
blk.11.ffn_norm.weightF32[1536]
-
blk.11.ffn_up.biasF32[6144]
-
blk.11.ffn_up.weightQ8_0[1536, 6144]
-
blk.12.attn_norm.biasF32[1536]
-
blk.12.attn_norm.weightF32[1536]
-
blk.12.attn_output.biasF32[1536]
-
blk.12.attn_output.weightQ8_0[1536, 1536]
-
blk.12.attn_qkv.biasF32[4608]
-
blk.12.attn_qkv.weightQ8_0[1536, 4608]
-
blk.12.ffn_down.biasF32[1536]
-
blk.12.ffn_down.weightQ8_0[6144, 1536]
-
blk.12.ffn_norm.biasF32[1536]
-
blk.12.ffn_norm.weightF32[1536]
-
blk.12.ffn_up.biasF32[6144]
-
blk.12.ffn_up.weightQ8_0[1536, 6144]
-
blk.13.attn_norm.biasF32[1536]
-
blk.13.attn_norm.weightF32[1536]
-
blk.13.attn_output.biasF32[1536]
-
blk.13.attn_output.weightQ8_0[1536, 1536]
-
blk.13.attn_qkv.biasF32[4608]
-
blk.13.attn_qkv.weightQ8_0[1536, 4608]
-
blk.13.ffn_down.biasF32[1536]
-
blk.13.ffn_down.weightQ8_0[6144, 1536]
-
blk.13.ffn_norm.biasF32[1536]
-
blk.13.ffn_norm.weightF32[1536]
-
blk.13.ffn_up.biasF32[6144]
-
blk.13.ffn_up.weightQ8_0[1536, 6144]
-
blk.14.attn_norm.biasF32[1536]
-
blk.14.attn_norm.weightF32[1536]
-
blk.14.attn_output.biasF32[1536]
-
blk.14.attn_output.weightQ8_0[1536, 1536]
-
blk.14.attn_qkv.biasF32[4608]
-
blk.14.attn_qkv.weightQ8_0[1536, 4608]
-
blk.14.ffn_down.biasF32[1536]
-
blk.14.ffn_down.weightQ8_0[6144, 1536]
-
blk.14.ffn_norm.biasF32[1536]
-
blk.14.ffn_norm.weightF32[1536]
-
blk.14.ffn_up.biasF32[6144]
-
blk.14.ffn_up.weightQ8_0[1536, 6144]
-
blk.15.attn_norm.biasF32[1536]
-
blk.15.attn_norm.weightF32[1536]
-
blk.15.attn_output.biasF32[1536]
-
blk.15.attn_output.weightQ8_0[1536, 1536]
-
blk.15.attn_qkv.biasF32[4608]
-
blk.15.attn_qkv.weightQ8_0[1536, 4608]
-
blk.15.ffn_down.biasF32[1536]
-
blk.15.ffn_down.weightQ8_0[6144, 1536]
-
blk.15.ffn_norm.biasF32[1536]
-
blk.15.ffn_norm.weightF32[1536]
-
blk.15.ffn_up.biasF32[6144]
-
blk.15.ffn_up.weightQ8_0[1536, 6144]
-
blk.16.attn_norm.biasF32[1536]
-
blk.16.attn_norm.weightF32[1536]
-
blk.16.attn_output.biasF32[1536]
-
blk.16.attn_output.weightQ8_0[1536, 1536]
-
blk.16.attn_qkv.biasF32[4608]
-
blk.16.attn_qkv.weightQ8_0[1536, 4608]
-
blk.16.ffn_down.biasF32[1536]
-
blk.16.ffn_down.weightQ8_0[6144, 1536]
-
blk.16.ffn_norm.biasF32[1536]
-
blk.16.ffn_norm.weightF32[1536]
-
blk.16.ffn_up.biasF32[6144]
-
blk.16.ffn_up.weightQ8_0[1536, 6144]
-
blk.17.attn_norm.biasF32[1536]
-
blk.17.attn_norm.weightF32[1536]
-
blk.17.attn_output.biasF32[1536]
-
blk.17.attn_output.weightQ8_0[1536, 1536]
-
blk.17.attn_qkv.biasF32[4608]
-
blk.17.attn_qkv.weightQ8_0[1536, 4608]
-
blk.17.ffn_down.biasF32[1536]
-
blk.17.ffn_down.weightQ8_0[6144, 1536]
-
blk.17.ffn_norm.biasF32[1536]
-
blk.17.ffn_norm.weightF32[1536]
-
blk.17.ffn_up.biasF32[6144]
-
blk.17.ffn_up.weightQ8_0[1536, 6144]
-
blk.18.attn_norm.biasF32[1536]
-
blk.18.attn_norm.weightF32[1536]
-
blk.18.attn_output.biasF32[1536]
-
blk.18.attn_output.weightQ8_0[1536, 1536]
-
blk.18.attn_qkv.biasF32[4608]
-
blk.18.attn_qkv.weightQ8_0[1536, 4608]
-
blk.18.ffn_down.biasF32[1536]
-
blk.18.ffn_down.weightQ8_0[6144, 1536]
-
blk.18.ffn_norm.biasF32[1536]
-
blk.18.ffn_norm.weightF32[1536]
-
blk.18.ffn_up.biasF32[6144]
-
blk.18.ffn_up.weightQ8_0[1536, 6144]
-
blk.19.attn_norm.biasF32[1536]
-
blk.19.attn_norm.weightF32[1536]
-
blk.19.attn_output.biasF32[1536]
-
blk.19.attn_output.weightQ8_0[1536, 1536]
-
blk.19.attn_qkv.biasF32[4608]
-
blk.19.attn_qkv.weightQ8_0[1536, 4608]
-
blk.19.ffn_down.biasF32[1536]
-
blk.19.ffn_down.weightQ8_0[6144, 1536]
-
blk.19.ffn_norm.biasF32[1536]
-
blk.19.ffn_norm.weightF32[1536]
-
blk.19.ffn_up.biasF32[6144]
-
blk.19.ffn_up.weightQ8_0[1536, 6144]
-
blk.20.attn_norm.biasF32[1536]
-
blk.20.attn_norm.weightF32[1536]
-
blk.20.attn_output.biasF32[1536]
-
blk.20.attn_output.weightQ8_0[1536, 1536]
-
blk.20.attn_qkv.biasF32[4608]
-
blk.20.attn_qkv.weightQ8_0[1536, 4608]
-
blk.20.ffn_down.biasF32[1536]
-
blk.20.ffn_down.weightQ8_0[6144, 1536]
-
blk.20.ffn_norm.biasF32[1536]
-
blk.20.ffn_norm.weightF32[1536]
-
blk.20.ffn_up.biasF32[6144]
-
blk.20.ffn_up.weightQ8_0[1536, 6144]
-
blk.21.attn_norm.biasF32[1536]
-
blk.21.attn_norm.weightF32[1536]
-
blk.21.attn_output.biasF32[1536]
-
blk.21.attn_output.weightQ8_0[1536, 1536]
-
blk.21.attn_qkv.biasF32[4608]
-
blk.21.attn_qkv.weightQ8_0[1536, 4608]
-
blk.21.ffn_down.biasF32[1536]
-
blk.21.ffn_down.weightQ8_0[6144, 1536]
-
blk.21.ffn_norm.biasF32[1536]
-
blk.21.ffn_norm.weightF32[1536]
-
blk.21.ffn_up.biasF32[6144]
-
blk.21.ffn_up.weightQ8_0[1536, 6144]
-
blk.22.attn_norm.biasF32[1536]
-
blk.22.attn_norm.weightF32[1536]
-
blk.22.attn_output.biasF32[1536]
-
blk.22.attn_output.weightQ8_0[1536, 1536]
-
blk.22.attn_qkv.biasF32[4608]
-
blk.22.attn_qkv.weightQ8_0[1536, 4608]
-
blk.22.ffn_down.biasF32[1536]
-
blk.22.ffn_down.weightQ8_0[6144, 1536]
-
blk.22.ffn_norm.biasF32[1536]
-
blk.22.ffn_norm.weightF32[1536]
-
blk.22.ffn_up.biasF32[6144]
-
blk.22.ffn_up.weightQ8_0[1536, 6144]
-
blk.23.attn_norm.biasF32[1536]
-
blk.23.attn_norm.weightF32[1536]
-
blk.23.attn_output.biasF32[1536]
-
blk.23.attn_output.weightQ8_0[1536, 1536]
-
blk.23.attn_qkv.biasF32[4608]
-
blk.23.attn_qkv.weightQ8_0[1536, 4608]
-
blk.23.ffn_down.biasF32[1536]
-
blk.23.ffn_down.weightQ8_0[6144, 1536]
-
blk.23.ffn_norm.biasF32[1536]
-
blk.23.ffn_norm.weightF32[1536]
-
blk.23.ffn_up.biasF32[6144]
-
blk.23.ffn_up.weightQ8_0[1536, 6144]
-
output.weightQ8_0[1536, 50257]
-
output_norm.biasF32[1536]
-
position_embd.weightF32[1536, 2048]
-
output_norm.weightF32[1536]
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
blk.12
blk.13
blk.14
blk.15
blk.16
blk.17
blk.18
blk.19
blk.20
blk.21
blk.22
blk.23