Salamandra-7b model ready to use with Ollama it can response in Spanish, Catalan, Galician and Basque
198 Pulls Updated 7 weeks ago
da90a4badcb4 · 4.9GB
-
general.architecturellamallama
-
general.base_model.0.nameSalamandra 7bSalamandra 7b
-
general.base_model.0.organizationBSC LTBSC LT
-
general.base_model.0.repo_urlhttps://huggingface.co/BSC-LT/salamandra-7bhttps://huggingface.co/BSC-LT/salamandra-7b
-
general.base_model.count11
-
general.basenamesalamandrasalamandra
-
general.dataset.0.nameColossal Oscar 1.0Colossal Oscar 1.0
-
general.dataset.0.organizationOscar CorpusOscar Corpus
-
general.dataset.0.repo_urlhttps://huggingface.co/oscar-corpus/colossal-oscar-1.0https://huggingface.co/oscar-corpus/colossal-oscar-1.0
-
general.dataset.0.version1.01.0
-
general.dataset.1.nameFineweb EduFineweb Edu
-
general.dataset.1.organizationHuggingFaceFWHuggingFaceFW
-
general.dataset.1.repo_urlhttps://huggingface.co/HuggingFaceFW/fineweb-eduhttps://huggingface.co/HuggingFaceFW/fineweb-edu
-
general.dataset.10.nameFrench PD BooksFrench PD Books
-
general.dataset.10.organizationPleIAsPleIAs
-
general.dataset.10.repo_urlhttps://huggingface.co/PleIAs/French-PD-Bookshttps://huggingface.co/PleIAs/French-PD-Books
-
general.dataset.11.nameGreek_Legal_CodeGreek_Legal_Code
-
general.dataset.11.organizationAI Team UoAAI Team UoA
-
general.dataset.11.repo_urlhttps://huggingface.co/AI-team-UoA/greek_legal_codehttps://huggingface.co/AI-team-UoA/greek_legal_code
-
general.dataset.12.nameLatxa Corpus v1.1Latxa Corpus v1.1
-
general.dataset.12.organizationHiTZHiTZ
-
general.dataset.12.repo_urlhttps://huggingface.co/HiTZ/latxa-corpus-v1.1https://huggingface.co/HiTZ/latxa-corpus-v1.1
-
general.dataset.12.versionv1.1v1.1
-
general.dataset.13.namepeS2opeS2o
-
general.dataset.13.organizationAllenaiAllenai
-
general.dataset.13.repo_urlhttps://huggingface.co/allenai/peS2ohttps://huggingface.co/allenai/peS2o
-
general.dataset.14.namePile Of LawPile Of Law
-
general.dataset.14.organizationPile Of LawPile Of Law
-
general.dataset.14.repo_urlhttps://huggingface.co/pile-of-law/pile-of-lawhttps://huggingface.co/pile-of-law/pile-of-law
-
general.dataset.15.nameParlamento PtParlamento Pt
-
general.dataset.15.organizationPORTULANPORTULAN
-
general.dataset.15.repo_urlhttps://huggingface.co/PORTULAN/parlamento-pthttps://huggingface.co/PORTULAN/parlamento-pt
-
general.dataset.16.nameProof PileProof Pile
-
general.dataset.16.organizationHoskinson CenterHoskinson Center
-
general.dataset.16.repo_urlhttps://huggingface.co/hoskinson-center/proof-pilehttps://huggingface.co/hoskinson-center/proof-pile
-
general.dataset.17.nameRedPajama Data 1TRedPajama Data 1T
-
general.dataset.17.organizationTogethercomputerTogethercomputer
-
general.dataset.17.repo_urlhttps://huggingface.co/togethercomputer/RedPajama-Data-1Thttps://huggingface.co/togethercomputer/RedPajama-Data-1T
-
general.dataset.18.nameStarcoderdataStarcoderdata
-
general.dataset.18.organizationBigcodeBigcode
-
general.dataset.18.repo_urlhttps://huggingface.co/bigcode/starcoderdatahttps://huggingface.co/bigcode/starcoderdata
-
general.dataset.19.nameTagesschau 2018 2023Tagesschau 2018 2023
-
general.dataset.19.organizationBjoernpBjoernp
-
general.dataset.19.repo_urlhttps://huggingface.co/bjoernp/tagesschau-2018-2023https://huggingface.co/bjoernp/tagesschau-2018-2023
-
general.dataset.19.version2018-20232018-2023
-
general.dataset.2.nameEurlex_ResourcesEurlex_Resources
-
general.dataset.2.organizationJoelniklausJoelniklaus
-
general.dataset.2.repo_urlhttps://huggingface.co/joelniklaus/eurlex_resourceshttps://huggingface.co/joelniklaus/eurlex_resources
-
general.dataset.20.nameThe_Pile_DeduplicatedThe_Pile_Deduplicated
-
general.dataset.20.organizationEleutherAIEleutherAI
-
general.dataset.20.repo_urlhttps://huggingface.co/EleutherAI/the_pile_deduplicatedhttps://huggingface.co/EleutherAI/the_pile_deduplicated
-
general.dataset.3.nameLegal Mc4Legal Mc4
-
general.dataset.3.organizationJoelitoJoelito
-
general.dataset.3.repo_urlhttps://huggingface.co/joelito/legal-mc4https://huggingface.co/joelito/legal-mc4
-
general.dataset.4.nameCATalogCATalog
-
general.dataset.4.organizationProjecte AinaProjecte Aina
-
general.dataset.4.repo_urlhttps://huggingface.co/projecte-aina/CATaloghttps://huggingface.co/projecte-aina/CATalog
-
general.dataset.5.nameBrwacBrwac
-
general.dataset.5.organizationUFRGSUFRGS
-
general.dataset.5.repo_urlhttps://huggingface.co/UFRGS/brwachttps://huggingface.co/UFRGS/brwac
-
general.dataset.6.nameHrwacHrwac
-
general.dataset.6.organizationCommunity DatasetsCommunity Datasets
-
general.dataset.6.repo_urlhttps://huggingface.co/community-datasets/hrwachttps://huggingface.co/community-datasets/hrwac
-
general.dataset.7.nameDanish GigawordDanish Gigaword
-
general.dataset.7.organizationDanish Foundation ModelsDanish Foundation Models
-
general.dataset.7.repo_urlhttps://huggingface.co/danish-foundation-models/danish-gigawordhttps://huggingface.co/danish-foundation-models/danish-gigaword
-
general.dataset.8.nameEuscrawlEuscrawl
-
general.dataset.8.organizationHiTZHiTZ
-
general.dataset.8.repo_urlhttps://huggingface.co/HiTZ/euscrawlhttps://huggingface.co/HiTZ/euscrawl
-
general.dataset.9.nameFrench PD NewspapersFrench PD Newspapers
-
general.dataset.9.organizationPleIAsPleIAs
-
general.dataset.9.repo_urlhttps://huggingface.co/PleIAs/French-PD-Newspapershttps://huggingface.co/PleIAs/French-PD-Newspapers
-
general.dataset.count2121
-
general.file_type1515
-
general.finetuneinstructinstruct
-
general.languages[bg, ca, code, cs, cy, ...][bg, ca, code, cs, cy, ...]
-
general.licenseapache-2.0apache-2.0
-
general.nameSalamandra 7b InstructSalamandra 7b Instruct
-
general.organizationBSC LTBSC LT
-
general.quantization_version22
-
general.size_label7B7B
-
general.tags[text-generation][text-generation]
-
general.typemodelmodel
-
llama.attention.head_count3232
-
llama.attention.head_count_kv88
-
llama.attention.key_length128128
-
llama.attention.layer_norm_rms_epsilon1e-051e-05
-
llama.attention.value_length128128
-
llama.block_count3232
-
llama.context_length81928192
-
llama.embedding_length40964096
-
llama.feed_forward_length1100811008
-
llama.rope.dimension_count128128
-
llama.rope.freq_base1000010000
-
llama.vocab_size256000256000
-
tokenizer.ggml.add_bos_tokentruetrue
-
tokenizer.ggml.add_eos_tokenfalsefalse
-
tokenizer.ggml.add_space_prefixtruetrue
-
tokenizer.ggml.bos_token_id11
-
tokenizer.ggml.eos_token_id22
-
tokenizer.ggml.modelllamallama
-
tokenizer.ggml.padding_token_id00
-
tokenizer.ggml.predefaultdefault
-
tokenizer.ggml.scores[-1000, -1000, -1000, 0, -1000, ...][-1000, -1000, -1000, 0, -1000, ...]
-
tokenizer.ggml.token_type[3, 3, 3, 3, 3, ...][3, 3, 3, 3, 3, ...]
-
tokenizer.ggml.tokens[<unk>, <s>, </s>, <pad>, <|im_start|>, ...][<unk>, <s>, </s>, <pad>, <|im_start|>, ...]
-
tokenizer.ggml.unknown_token_id00
-
token_embd.weightQ4_K[4096, 256000]
-
blk.0.attn_k.weightQ4_K[4096, 1024]
-
blk.0.attn_norm.weightF32[4096]
-
blk.0.attn_output.weightQ4_K[4096, 4096]
-
blk.0.attn_q.weightQ4_K[4096, 4096]
-
blk.0.attn_v.weightQ6_K[4096, 1024]
-
blk.0.ffn_down.weightQ6_K[11008, 4096]
-
blk.0.ffn_gate.weightQ4_K[4096, 11008]
-
blk.0.ffn_norm.weightF32[4096]
-
blk.0.ffn_up.weightQ4_K[4096, 11008]
-
blk.1.attn_k.weightQ4_K[4096, 1024]
-
blk.1.attn_norm.weightF32[4096]
-
blk.1.attn_output.weightQ4_K[4096, 4096]
-
blk.1.attn_q.weightQ4_K[4096, 4096]
-
blk.1.attn_v.weightQ6_K[4096, 1024]
-
blk.1.ffn_down.weightQ6_K[11008, 4096]
-
blk.1.ffn_gate.weightQ4_K[4096, 11008]
-
blk.1.ffn_norm.weightF32[4096]
-
blk.1.ffn_up.weightQ4_K[4096, 11008]
-
blk.2.attn_k.weightQ4_K[4096, 1024]
-
blk.2.attn_norm.weightF32[4096]
-
blk.2.attn_output.weightQ4_K[4096, 4096]
-
blk.2.attn_q.weightQ4_K[4096, 4096]
-
blk.2.attn_v.weightQ6_K[4096, 1024]
-
blk.2.ffn_down.weightQ6_K[11008, 4096]
-
blk.2.ffn_gate.weightQ4_K[4096, 11008]
-
blk.2.ffn_norm.weightF32[4096]
-
blk.2.ffn_up.weightQ4_K[4096, 11008]
-
blk.3.attn_k.weightQ4_K[4096, 1024]
-
blk.3.attn_norm.weightF32[4096]
-
blk.3.attn_output.weightQ4_K[4096, 4096]
-
blk.3.attn_q.weightQ4_K[4096, 4096]
-
blk.3.attn_v.weightQ6_K[4096, 1024]
-
blk.3.ffn_down.weightQ6_K[11008, 4096]
-
blk.3.ffn_gate.weightQ4_K[4096, 11008]
-
blk.3.ffn_norm.weightF32[4096]
-
blk.3.ffn_up.weightQ4_K[4096, 11008]
-
blk.4.attn_k.weightQ4_K[4096, 1024]
-
blk.4.attn_norm.weightF32[4096]
-
blk.4.attn_output.weightQ4_K[4096, 4096]
-
blk.4.attn_q.weightQ4_K[4096, 4096]
-
blk.4.attn_v.weightQ4_K[4096, 1024]
-
blk.4.ffn_down.weightQ4_K[11008, 4096]
-
blk.4.ffn_gate.weightQ4_K[4096, 11008]
-
blk.4.ffn_norm.weightF32[4096]
-
blk.4.ffn_up.weightQ4_K[4096, 11008]
-
blk.5.attn_k.weightQ4_K[4096, 1024]
-
blk.5.attn_norm.weightF32[4096]
-
blk.5.attn_output.weightQ4_K[4096, 4096]
-
blk.5.attn_q.weightQ4_K[4096, 4096]
-
blk.5.attn_v.weightQ4_K[4096, 1024]
-
blk.5.ffn_down.weightQ4_K[11008, 4096]
-
blk.5.ffn_gate.weightQ4_K[4096, 11008]
-
blk.5.ffn_norm.weightF32[4096]
-
blk.5.ffn_up.weightQ4_K[4096, 11008]
-
blk.6.attn_k.weightQ4_K[4096, 1024]
-
blk.6.attn_norm.weightF32[4096]
-
blk.6.attn_output.weightQ4_K[4096, 4096]
-
blk.6.attn_q.weightQ4_K[4096, 4096]
-
blk.6.attn_v.weightQ6_K[4096, 1024]
-
blk.6.ffn_down.weightQ6_K[11008, 4096]
-
blk.6.ffn_gate.weightQ4_K[4096, 11008]
-
blk.6.ffn_norm.weightF32[4096]
-
blk.6.ffn_up.weightQ4_K[4096, 11008]
-
blk.7.attn_k.weightQ4_K[4096, 1024]
-
blk.7.attn_norm.weightF32[4096]
-
blk.7.attn_output.weightQ4_K[4096, 4096]
-
blk.7.attn_q.weightQ4_K[4096, 4096]
-
blk.7.attn_v.weightQ4_K[4096, 1024]
-
blk.7.ffn_down.weightQ4_K[11008, 4096]
-
blk.7.ffn_gate.weightQ4_K[4096, 11008]
-
blk.7.ffn_norm.weightF32[4096]
-
blk.7.ffn_up.weightQ4_K[4096, 11008]
-
blk.8.attn_k.weightQ4_K[4096, 1024]
-
blk.8.attn_norm.weightF32[4096]
-
blk.8.attn_output.weightQ4_K[4096, 4096]
-
blk.8.attn_q.weightQ4_K[4096, 4096]
-
blk.8.attn_v.weightQ4_K[4096, 1024]
-
blk.8.ffn_down.weightQ4_K[11008, 4096]
-
blk.8.ffn_gate.weightQ4_K[4096, 11008]
-
blk.8.ffn_norm.weightF32[4096]
-
blk.8.ffn_up.weightQ4_K[4096, 11008]
-
blk.9.attn_k.weightQ4_K[4096, 1024]
-
blk.9.attn_norm.weightF32[4096]
-
blk.9.attn_output.weightQ4_K[4096, 4096]
-
blk.9.attn_q.weightQ4_K[4096, 4096]
-
blk.9.attn_v.weightQ4_K[4096, 1024]
-
blk.9.ffn_down.weightQ6_K[11008, 4096]
-
blk.9.ffn_gate.weightQ4_K[4096, 11008]
-
blk.9.ffn_norm.weightF32[4096]
-
blk.9.ffn_up.weightQ4_K[4096, 11008]
-
blk.10.attn_k.weightQ4_K[4096, 1024]
-
blk.10.attn_norm.weightF32[4096]
-
blk.10.attn_output.weightQ4_K[4096, 4096]
-
blk.10.attn_q.weightQ4_K[4096, 4096]
-
blk.10.attn_v.weightQ6_K[4096, 1024]
-
blk.10.ffn_down.weightQ4_K[11008, 4096]
-
blk.10.ffn_gate.weightQ4_K[4096, 11008]
-
blk.10.ffn_norm.weightF32[4096]
-
blk.10.ffn_up.weightQ4_K[4096, 11008]
-
blk.11.attn_k.weightQ4_K[4096, 1024]
-
blk.11.attn_norm.weightF32[4096]
-
blk.11.attn_output.weightQ4_K[4096, 4096]
-
blk.11.attn_q.weightQ4_K[4096, 4096]
-
blk.11.attn_v.weightQ4_K[4096, 1024]
-
blk.11.ffn_down.weightQ6_K[11008, 4096]
-
blk.11.ffn_gate.weightQ4_K[4096, 11008]
-
blk.11.ffn_norm.weightF32[4096]
-
blk.11.ffn_up.weightQ4_K[4096, 11008]
-
blk.12.attn_k.weightQ4_K[4096, 1024]
-
blk.12.attn_norm.weightF32[4096]
-
blk.12.attn_output.weightQ4_K[4096, 4096]
-
blk.12.attn_q.weightQ4_K[4096, 4096]
-
blk.12.attn_v.weightQ4_K[4096, 1024]
-
blk.12.ffn_down.weightQ4_K[11008, 4096]
-
blk.12.ffn_gate.weightQ4_K[4096, 11008]
-
blk.12.ffn_norm.weightF32[4096]
-
blk.12.ffn_up.weightQ4_K[4096, 11008]
-
blk.13.attn_k.weightQ4_K[4096, 1024]
-
blk.13.attn_norm.weightF32[4096]
-
blk.13.attn_output.weightQ4_K[4096, 4096]
-
blk.13.attn_q.weightQ4_K[4096, 4096]
-
blk.13.attn_v.weightQ6_K[4096, 1024]
-
blk.13.ffn_down.weightQ4_K[11008, 4096]
-
blk.13.ffn_gate.weightQ4_K[4096, 11008]
-
blk.13.ffn_norm.weightF32[4096]
-
blk.13.ffn_up.weightQ4_K[4096, 11008]
-
blk.14.attn_k.weightQ4_K[4096, 1024]
-
blk.14.attn_norm.weightF32[4096]
-
blk.14.attn_output.weightQ4_K[4096, 4096]
-
blk.14.attn_q.weightQ4_K[4096, 4096]
-
blk.14.attn_v.weightQ4_K[4096, 1024]
-
blk.14.ffn_down.weightQ6_K[11008, 4096]
-
blk.14.ffn_gate.weightQ4_K[4096, 11008]
-
blk.14.ffn_norm.weightF32[4096]
-
blk.14.ffn_up.weightQ4_K[4096, 11008]
-
blk.15.attn_k.weightQ4_K[4096, 1024]
-
blk.15.attn_norm.weightF32[4096]
-
blk.15.attn_output.weightQ4_K[4096, 4096]
-
blk.15.attn_q.weightQ4_K[4096, 4096]
-
blk.15.attn_v.weightQ4_K[4096, 1024]
-
blk.15.ffn_down.weightQ4_K[11008, 4096]
-
blk.15.ffn_gate.weightQ4_K[4096, 11008]
-
blk.15.ffn_norm.weightF32[4096]
-
blk.15.ffn_up.weightQ4_K[4096, 11008]
-
blk.16.attn_k.weightQ4_K[4096, 1024]
-
blk.16.attn_norm.weightF32[4096]
-
blk.16.attn_output.weightQ4_K[4096, 4096]
-
blk.16.attn_q.weightQ4_K[4096, 4096]
-
blk.16.attn_v.weightQ6_K[4096, 1024]
-
blk.16.ffn_down.weightQ4_K[11008, 4096]
-
blk.16.ffn_gate.weightQ4_K[4096, 11008]
-
blk.16.ffn_norm.weightF32[4096]
-
blk.16.ffn_up.weightQ4_K[4096, 11008]
-
blk.17.attn_k.weightQ4_K[4096, 1024]
-
blk.17.attn_norm.weightF32[4096]
-
blk.17.attn_output.weightQ4_K[4096, 4096]
-
blk.17.attn_q.weightQ4_K[4096, 4096]
-
blk.17.attn_v.weightQ4_K[4096, 1024]
-
blk.17.ffn_down.weightQ6_K[11008, 4096]
-
blk.17.ffn_gate.weightQ4_K[4096, 11008]
-
blk.17.ffn_norm.weightF32[4096]
-
blk.17.ffn_up.weightQ4_K[4096, 11008]
-
blk.18.attn_k.weightQ4_K[4096, 1024]
-
blk.18.attn_norm.weightF32[4096]
-
blk.18.attn_output.weightQ4_K[4096, 4096]
-
blk.18.attn_q.weightQ4_K[4096, 4096]
-
blk.18.attn_v.weightQ4_K[4096, 1024]
-
blk.18.ffn_down.weightQ4_K[11008, 4096]
-
blk.18.ffn_gate.weightQ4_K[4096, 11008]
-
blk.18.ffn_norm.weightF32[4096]
-
blk.18.ffn_up.weightQ4_K[4096, 11008]
-
blk.19.attn_k.weightQ4_K[4096, 1024]
-
blk.19.attn_norm.weightF32[4096]
-
blk.19.attn_output.weightQ4_K[4096, 4096]
-
blk.19.attn_q.weightQ4_K[4096, 4096]
-
blk.19.attn_v.weightQ6_K[4096, 1024]
-
blk.19.ffn_down.weightQ4_K[11008, 4096]
-
blk.19.ffn_gate.weightQ4_K[4096, 11008]
-
blk.19.ffn_norm.weightF32[4096]
-
blk.19.ffn_up.weightQ4_K[4096, 11008]
-
blk.20.attn_k.weightQ4_K[4096, 1024]
-
blk.20.attn_norm.weightF32[4096]
-
blk.20.attn_output.weightQ4_K[4096, 4096]
-
blk.20.attn_q.weightQ4_K[4096, 4096]
-
blk.20.attn_v.weightQ4_K[4096, 1024]
-
blk.20.ffn_down.weightQ6_K[11008, 4096]
-
blk.20.ffn_gate.weightQ4_K[4096, 11008]
-
blk.20.ffn_norm.weightF32[4096]
-
blk.20.ffn_up.weightQ4_K[4096, 11008]
-
blk.21.attn_k.weightQ4_K[4096, 1024]
-
blk.21.attn_norm.weightF32[4096]
-
blk.21.attn_output.weightQ4_K[4096, 4096]
-
blk.21.attn_q.weightQ4_K[4096, 4096]
-
blk.21.attn_v.weightQ4_K[4096, 1024]
-
blk.21.ffn_down.weightQ4_K[11008, 4096]
-
blk.21.ffn_gate.weightQ4_K[4096, 11008]
-
blk.21.ffn_norm.weightF32[4096]
-
blk.21.ffn_up.weightQ4_K[4096, 11008]
-
blk.22.attn_k.weightQ4_K[4096, 1024]
-
blk.22.attn_norm.weightF32[4096]
-
blk.22.attn_output.weightQ4_K[4096, 4096]
-
blk.22.attn_q.weightQ4_K[4096, 4096]
-
blk.22.attn_v.weightQ6_K[4096, 1024]
-
blk.22.ffn_down.weightQ4_K[11008, 4096]
-
blk.22.ffn_gate.weightQ4_K[4096, 11008]
-
blk.22.ffn_norm.weightF32[4096]
-
blk.22.ffn_up.weightQ4_K[4096, 11008]
-
blk.23.attn_k.weightQ4_K[4096, 1024]
-
blk.23.attn_norm.weightF32[4096]
-
blk.23.attn_output.weightQ4_K[4096, 4096]
-
blk.23.attn_q.weightQ4_K[4096, 4096]
-
blk.23.attn_v.weightQ4_K[4096, 1024]
-
blk.23.ffn_down.weightQ4_K[11008, 4096]
-
blk.23.ffn_gate.weightQ4_K[4096, 11008]
-
blk.23.ffn_norm.weightF32[4096]
-
blk.23.ffn_up.weightQ4_K[4096, 11008]
-
blk.24.attn_k.weightQ4_K[4096, 1024]
-
blk.24.attn_norm.weightF32[4096]
-
blk.24.attn_output.weightQ4_K[4096, 4096]
-
blk.24.attn_q.weightQ4_K[4096, 4096]
-
blk.24.attn_v.weightQ6_K[4096, 1024]
-
blk.24.ffn_down.weightQ6_K[11008, 4096]
-
blk.24.ffn_gate.weightQ4_K[4096, 11008]
-
blk.24.ffn_norm.weightF32[4096]
-
blk.24.ffn_up.weightQ4_K[4096, 11008]
-
blk.25.attn_k.weightQ4_K[4096, 1024]
-
blk.25.attn_norm.weightF32[4096]
-
blk.25.attn_output.weightQ4_K[4096, 4096]
-
blk.25.attn_q.weightQ4_K[4096, 4096]
-
blk.25.attn_v.weightQ4_K[4096, 1024]
-
blk.25.ffn_down.weightQ4_K[11008, 4096]
-
blk.25.ffn_gate.weightQ4_K[4096, 11008]
-
blk.25.ffn_norm.weightF32[4096]
-
blk.25.ffn_up.weightQ4_K[4096, 11008]
-
blk.26.attn_k.weightQ4_K[4096, 1024]
-
blk.26.attn_norm.weightF32[4096]
-
blk.26.attn_output.weightQ4_K[4096, 4096]
-
blk.26.attn_q.weightQ4_K[4096, 4096]
-
blk.26.attn_v.weightQ4_K[4096, 1024]
-
blk.26.ffn_down.weightQ4_K[11008, 4096]
-
blk.26.ffn_gate.weightQ4_K[4096, 11008]
-
blk.26.ffn_norm.weightF32[4096]
-
blk.26.ffn_up.weightQ4_K[4096, 11008]
-
blk.27.attn_k.weightQ4_K[4096, 1024]
-
blk.27.attn_norm.weightF32[4096]
-
blk.27.attn_output.weightQ4_K[4096, 4096]
-
blk.27.attn_q.weightQ4_K[4096, 4096]
-
blk.27.attn_v.weightQ6_K[4096, 1024]
-
blk.27.ffn_down.weightQ6_K[11008, 4096]
-
blk.27.ffn_gate.weightQ4_K[4096, 11008]
-
blk.27.ffn_norm.weightF32[4096]
-
blk.27.ffn_up.weightQ4_K[4096, 11008]
-
blk.28.attn_k.weightQ4_K[4096, 1024]
-
blk.28.attn_norm.weightF32[4096]
-
blk.28.attn_output.weightQ4_K[4096, 4096]
-
blk.28.attn_q.weightQ4_K[4096, 4096]
-
blk.28.attn_v.weightQ6_K[4096, 1024]
-
blk.28.ffn_down.weightQ6_K[11008, 4096]
-
blk.28.ffn_gate.weightQ4_K[4096, 11008]
-
blk.28.ffn_norm.weightF32[4096]
-
blk.28.ffn_up.weightQ4_K[4096, 11008]
-
blk.29.attn_k.weightQ4_K[4096, 1024]
-
blk.29.attn_norm.weightF32[4096]
-
blk.29.attn_output.weightQ4_K[4096, 4096]
-
blk.29.attn_q.weightQ4_K[4096, 4096]
-
blk.29.attn_v.weightQ6_K[4096, 1024]
-
blk.29.ffn_down.weightQ6_K[11008, 4096]
-
blk.29.ffn_gate.weightQ4_K[4096, 11008]
-
blk.29.ffn_norm.weightF32[4096]
-
blk.29.ffn_up.weightQ4_K[4096, 11008]
-
blk.30.attn_k.weightQ4_K[4096, 1024]
-
blk.30.attn_norm.weightF32[4096]
-
blk.30.attn_output.weightQ4_K[4096, 4096]
-
blk.30.attn_q.weightQ4_K[4096, 4096]
-
blk.30.attn_v.weightQ6_K[4096, 1024]
-
blk.30.ffn_down.weightQ6_K[11008, 4096]
-
blk.30.ffn_gate.weightQ4_K[4096, 11008]
-
blk.30.ffn_norm.weightF32[4096]
-
blk.30.ffn_up.weightQ4_K[4096, 11008]
-
blk.31.attn_k.weightQ4_K[4096, 1024]
-
blk.31.attn_norm.weightF32[4096]
-
blk.31.attn_output.weightQ4_K[4096, 4096]
-
blk.31.attn_q.weightQ4_K[4096, 4096]
-
blk.31.attn_v.weightQ6_K[4096, 1024]
-
blk.31.ffn_down.weightQ6_K[11008, 4096]
-
blk.31.ffn_gate.weightQ4_K[4096, 11008]
-
blk.31.ffn_norm.weightF32[4096]
-
blk.31.ffn_up.weightQ4_K[4096, 11008]
-
output.weightQ6_K[4096, 256000]
-
output_norm.weightF32[4096]
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
blk.24
blk.25
blk.26
blk.27
blk.28
blk.29
blk.30
blk.31