Uncensored, 8x7b and 8x22b fine-tuned models based on the Mixtral mixture of experts models that excels at coding tasks. Created by Eric Hartford.
8x7b
8x22b
455.9K Pulls Updated yesterday
5e241ec72b81 · 27GB
-
general.architecturellama
-
general.basenamedolphin-2.6-mixtral
-
general.dataset.0.nameDolphin
-
general.dataset.0.organizationEhartford
-
general.dataset.0.repo_urlhttps://huggingface.co/ehartford/dolphin
-
general.dataset.1.nameAiroboros 2.2.1
-
general.dataset.1.organizationJondurbin
-
general.dataset.1.repo_urlhttps://huggingface.co/jondurbin/airoboros-2.2.1
-
general.dataset.1.version2.2.1
-
general.dataset.2.nameDolphin Coder
-
general.dataset.2.organizationEhartford
-
general.dataset.2.repo_urlhttps://huggingface.co/ehartford/dolphin-coder
-
general.dataset.3.nameOpenhermes
-
general.dataset.3.organizationTeknium
-
general.dataset.3.repo_urlhttps://huggingface.co/teknium/openhermes
-
general.dataset.4.nameMagicoder OSS Instruct 75K
-
general.dataset.4.organizationIse Uiuc
-
general.dataset.4.repo_urlhttps://huggingface.co/ise-uiuc/Magicoder-OSS-Instruct-75K
-
general.dataset.5.nameMagicoder Evol Instruct 110K
-
general.dataset.5.organizationIse Uiuc
-
general.dataset.5.repo_urlhttps://huggingface.co/ise-uiuc/Magicoder-Evol-Instruct-110K
-
general.dataset.6.nameCapybara
-
general.dataset.6.organizationLDJnr
-
general.dataset.6.repo_urlhttps://huggingface.co/LDJnr/Capybara
-
general.dataset.count7
-
general.file_type14
-
general.languages[en]
-
general.licenseapache-2.0
-
general.nameDolphin 2.6 Mixtral 8x7b
-
general.quantization_version2
-
general.size_label8x7B
-
general.typemodel
-
llama.attention.head_count32
-
llama.attention.head_count_kv8
-
llama.attention.layer_norm_rms_epsilon1e-05
-
llama.block_count32
-
llama.context_length32768
-
llama.embedding_length4096
-
llama.expert_count8
-
llama.expert_used_count2
-
llama.feed_forward_length14336
-
llama.rope.dimension_count128
-
llama.rope.freq_base1e+06
-
llama.vocab_size32002
-
tokenizer.ggml.add_bos_tokentrue
-
tokenizer.ggml.add_eos_tokenfalse
-
tokenizer.ggml.bos_token_id1
-
tokenizer.ggml.eos_token_id32000
-
tokenizer.ggml.modelllama
-
tokenizer.ggml.predefault
-
tokenizer.ggml.scores[-1000, -1000, -1000, 0, 0, ...]
-
tokenizer.ggml.token_type[3, 3, 3, 6, 6, ...]
-
tokenizer.ggml.tokens[<unk>, <s>, </s>, <0x00>, <0x01>, ...]
-
NameTypeShape
-
token_embd.weightQ4_K[4096, 32002]
-
blk.0.attn_k.weightQ8_0[4096, 1024]
-
blk.0.attn_norm.weightF32[4096]
-
blk.0.attn_output.weightQ5_K[4096, 4096]
-
blk.0.attn_q.weightQ4_K[4096, 4096]
-
blk.0.attn_v.weightQ8_0[4096, 1024]
-
blk.0.ffn_down_exps.weightQ5_K[14336, 4096, 8]
-
blk.0.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.0.ffn_gate_inp.weightF32[4096, 8]
-
blk.0.ffn_norm.weightF32[4096]
-
blk.0.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.1.attn_k.weightQ8_0[4096, 1024]
-
blk.1.attn_norm.weightF32[4096]
-
blk.1.attn_output.weightQ5_K[4096, 4096]
-
blk.1.attn_q.weightQ4_K[4096, 4096]
-
blk.1.attn_v.weightQ8_0[4096, 1024]
-
blk.1.ffn_down_exps.weightQ5_K[14336, 4096, 8]
-
blk.1.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.1.ffn_gate_inp.weightF32[4096, 8]
-
blk.1.ffn_norm.weightF32[4096]
-
blk.1.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.2.attn_k.weightQ8_0[4096, 1024]
-
blk.2.attn_norm.weightF32[4096]
-
blk.2.attn_output.weightQ5_K[4096, 4096]
-
blk.2.attn_q.weightQ4_K[4096, 4096]
-
blk.2.attn_v.weightQ8_0[4096, 1024]
-
blk.2.ffn_down_exps.weightQ5_K[14336, 4096, 8]
-
blk.2.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.2.ffn_gate_inp.weightF32[4096, 8]
-
blk.2.ffn_norm.weightF32[4096]
-
blk.2.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.3.attn_k.weightQ8_0[4096, 1024]
-
blk.3.attn_norm.weightF32[4096]
-
blk.3.attn_output.weightQ5_K[4096, 4096]
-
blk.3.attn_q.weightQ4_K[4096, 4096]
-
blk.3.attn_v.weightQ8_0[4096, 1024]
-
blk.3.ffn_down_exps.weightQ5_K[14336, 4096, 8]
-
blk.3.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.3.ffn_gate_inp.weightF32[4096, 8]
-
blk.3.ffn_norm.weightF32[4096]
-
blk.3.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.4.attn_k.weightQ8_0[4096, 1024]
-
blk.4.attn_norm.weightF32[4096]
-
blk.4.attn_output.weightQ5_K[4096, 4096]
-
blk.4.attn_q.weightQ4_K[4096, 4096]
-
blk.4.attn_v.weightQ8_0[4096, 1024]
-
blk.4.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.4.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.4.ffn_gate_inp.weightF32[4096, 8]
-
blk.4.ffn_norm.weightF32[4096]
-
blk.4.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.5.attn_k.weightQ8_0[4096, 1024]
-
blk.5.attn_norm.weightF32[4096]
-
blk.5.attn_output.weightQ5_K[4096, 4096]
-
blk.5.attn_q.weightQ4_K[4096, 4096]
-
blk.5.attn_v.weightQ8_0[4096, 1024]
-
blk.5.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.5.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.5.ffn_gate_inp.weightF32[4096, 8]
-
blk.5.ffn_norm.weightF32[4096]
-
blk.5.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.6.attn_k.weightQ8_0[4096, 1024]
-
blk.6.attn_norm.weightF32[4096]
-
blk.6.attn_output.weightQ5_K[4096, 4096]
-
blk.6.attn_q.weightQ4_K[4096, 4096]
-
blk.6.attn_v.weightQ8_0[4096, 1024]
-
blk.6.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.6.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.6.ffn_gate_inp.weightF32[4096, 8]
-
blk.6.ffn_norm.weightF32[4096]
-
blk.6.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.7.attn_k.weightQ8_0[4096, 1024]
-
blk.7.attn_norm.weightF32[4096]
-
blk.7.attn_output.weightQ5_K[4096, 4096]
-
blk.7.attn_q.weightQ4_K[4096, 4096]
-
blk.7.attn_v.weightQ8_0[4096, 1024]
-
blk.7.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.7.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.7.ffn_gate_inp.weightF32[4096, 8]
-
blk.7.ffn_norm.weightF32[4096]
-
blk.7.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.8.attn_k.weightQ8_0[4096, 1024]
-
blk.8.attn_norm.weightF32[4096]
-
blk.8.attn_output.weightQ5_K[4096, 4096]
-
blk.8.attn_q.weightQ4_K[4096, 4096]
-
blk.8.attn_v.weightQ8_0[4096, 1024]
-
blk.8.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.8.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.8.ffn_gate_inp.weightF32[4096, 8]
-
blk.8.ffn_norm.weightF32[4096]
-
blk.8.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.9.attn_k.weightQ8_0[4096, 1024]
-
blk.9.attn_norm.weightF32[4096]
-
blk.9.attn_output.weightQ5_K[4096, 4096]
-
blk.9.attn_q.weightQ4_K[4096, 4096]
-
blk.9.attn_v.weightQ8_0[4096, 1024]
-
blk.9.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.9.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.9.ffn_gate_inp.weightF32[4096, 8]
-
blk.9.ffn_norm.weightF32[4096]
-
blk.9.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.10.attn_k.weightQ8_0[4096, 1024]
-
blk.10.attn_norm.weightF32[4096]
-
blk.10.attn_output.weightQ5_K[4096, 4096]
-
blk.10.attn_q.weightQ4_K[4096, 4096]
-
blk.10.attn_v.weightQ8_0[4096, 1024]
-
blk.10.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.10.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.10.ffn_gate_inp.weightF32[4096, 8]
-
blk.10.ffn_norm.weightF32[4096]
-
blk.10.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.11.attn_k.weightQ8_0[4096, 1024]
-
blk.11.attn_norm.weightF32[4096]
-
blk.11.attn_output.weightQ5_K[4096, 4096]
-
blk.11.attn_q.weightQ4_K[4096, 4096]
-
blk.11.attn_v.weightQ8_0[4096, 1024]
-
blk.11.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.11.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.11.ffn_gate_inp.weightF32[4096, 8]
-
blk.11.ffn_norm.weightF32[4096]
-
blk.11.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.12.attn_k.weightQ8_0[4096, 1024]
-
blk.12.attn_norm.weightF32[4096]
-
blk.12.attn_output.weightQ5_K[4096, 4096]
-
blk.12.attn_q.weightQ4_K[4096, 4096]
-
blk.12.attn_v.weightQ8_0[4096, 1024]
-
blk.12.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.12.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.12.ffn_gate_inp.weightF32[4096, 8]
-
blk.12.ffn_norm.weightF32[4096]
-
blk.12.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.13.attn_k.weightQ8_0[4096, 1024]
-
blk.13.attn_norm.weightF32[4096]
-
blk.13.attn_output.weightQ5_K[4096, 4096]
-
blk.13.attn_q.weightQ4_K[4096, 4096]
-
blk.13.attn_v.weightQ8_0[4096, 1024]
-
blk.13.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.13.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.13.ffn_gate_inp.weightF32[4096, 8]
-
blk.13.ffn_norm.weightF32[4096]
-
blk.13.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.14.attn_k.weightQ8_0[4096, 1024]
-
blk.14.attn_norm.weightF32[4096]
-
blk.14.attn_output.weightQ5_K[4096, 4096]
-
blk.14.attn_q.weightQ4_K[4096, 4096]
-
blk.14.attn_v.weightQ8_0[4096, 1024]
-
blk.14.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.14.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.14.ffn_gate_inp.weightF32[4096, 8]
-
blk.14.ffn_norm.weightF32[4096]
-
blk.14.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.15.attn_k.weightQ8_0[4096, 1024]
-
blk.15.attn_norm.weightF32[4096]
-
blk.15.attn_output.weightQ5_K[4096, 4096]
-
blk.15.attn_q.weightQ4_K[4096, 4096]
-
blk.15.attn_v.weightQ8_0[4096, 1024]
-
blk.15.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.15.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.15.ffn_gate_inp.weightF32[4096, 8]
-
blk.15.ffn_norm.weightF32[4096]
-
blk.15.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.16.attn_k.weightQ8_0[4096, 1024]
-
blk.16.attn_norm.weightF32[4096]
-
blk.16.attn_output.weightQ5_K[4096, 4096]
-
blk.16.attn_q.weightQ4_K[4096, 4096]
-
blk.16.attn_v.weightQ8_0[4096, 1024]
-
blk.16.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.16.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.16.ffn_gate_inp.weightF32[4096, 8]
-
blk.16.ffn_norm.weightF32[4096]
-
blk.16.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.17.attn_k.weightQ8_0[4096, 1024]
-
blk.17.attn_norm.weightF32[4096]
-
blk.17.attn_output.weightQ5_K[4096, 4096]
-
blk.17.attn_q.weightQ4_K[4096, 4096]
-
blk.17.attn_v.weightQ8_0[4096, 1024]
-
blk.17.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.17.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.17.ffn_gate_inp.weightF32[4096, 8]
-
blk.17.ffn_norm.weightF32[4096]
-
blk.17.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.18.attn_k.weightQ8_0[4096, 1024]
-
blk.18.attn_norm.weightF32[4096]
-
blk.18.attn_output.weightQ5_K[4096, 4096]
-
blk.18.attn_q.weightQ4_K[4096, 4096]
-
blk.18.attn_v.weightQ8_0[4096, 1024]
-
blk.18.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.18.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.18.ffn_gate_inp.weightF32[4096, 8]
-
blk.18.ffn_norm.weightF32[4096]
-
blk.18.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.19.attn_k.weightQ8_0[4096, 1024]
-
blk.19.attn_norm.weightF32[4096]
-
blk.19.attn_output.weightQ5_K[4096, 4096]
-
blk.19.attn_q.weightQ4_K[4096, 4096]
-
blk.19.attn_v.weightQ8_0[4096, 1024]
-
blk.19.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.19.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.19.ffn_gate_inp.weightF32[4096, 8]
-
blk.19.ffn_norm.weightF32[4096]
-
blk.19.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.20.attn_k.weightQ8_0[4096, 1024]
-
blk.20.attn_norm.weightF32[4096]
-
blk.20.attn_output.weightQ5_K[4096, 4096]
-
blk.20.attn_q.weightQ4_K[4096, 4096]
-
blk.20.attn_v.weightQ8_0[4096, 1024]
-
blk.20.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.20.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.20.ffn_gate_inp.weightF32[4096, 8]
-
blk.20.ffn_norm.weightF32[4096]
-
blk.20.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.21.attn_k.weightQ8_0[4096, 1024]
-
blk.21.attn_norm.weightF32[4096]
-
blk.21.attn_output.weightQ5_K[4096, 4096]
-
blk.21.attn_q.weightQ4_K[4096, 4096]
-
blk.21.attn_v.weightQ8_0[4096, 1024]
-
blk.21.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.21.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.21.ffn_gate_inp.weightF32[4096, 8]
-
blk.21.ffn_norm.weightF32[4096]
-
blk.21.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.22.attn_k.weightQ8_0[4096, 1024]
-
blk.22.attn_norm.weightF32[4096]
-
blk.22.attn_output.weightQ5_K[4096, 4096]
-
blk.22.attn_q.weightQ4_K[4096, 4096]
-
blk.22.attn_v.weightQ8_0[4096, 1024]
-
blk.22.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.22.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.22.ffn_gate_inp.weightF32[4096, 8]
-
blk.22.ffn_norm.weightF32[4096]
-
blk.22.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.23.attn_k.weightQ8_0[4096, 1024]
-
blk.23.attn_norm.weightF32[4096]
-
blk.23.attn_output.weightQ5_K[4096, 4096]
-
blk.23.attn_q.weightQ4_K[4096, 4096]
-
blk.23.attn_v.weightQ8_0[4096, 1024]
-
blk.23.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.23.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.23.ffn_gate_inp.weightF32[4096, 8]
-
blk.23.ffn_norm.weightF32[4096]
-
blk.23.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.24.attn_k.weightQ8_0[4096, 1024]
-
blk.24.attn_norm.weightF32[4096]
-
blk.24.attn_output.weightQ5_K[4096, 4096]
-
blk.24.attn_q.weightQ4_K[4096, 4096]
-
blk.24.attn_v.weightQ8_0[4096, 1024]
-
blk.24.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.24.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.24.ffn_gate_inp.weightF32[4096, 8]
-
blk.24.ffn_norm.weightF32[4096]
-
blk.24.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.25.attn_k.weightQ8_0[4096, 1024]
-
blk.25.attn_norm.weightF32[4096]
-
blk.25.attn_output.weightQ5_K[4096, 4096]
-
blk.25.attn_q.weightQ4_K[4096, 4096]
-
blk.25.attn_v.weightQ8_0[4096, 1024]
-
blk.25.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.25.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.25.ffn_gate_inp.weightF32[4096, 8]
-
blk.25.ffn_norm.weightF32[4096]
-
blk.25.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.26.attn_k.weightQ8_0[4096, 1024]
-
blk.26.attn_norm.weightF32[4096]
-
blk.26.attn_output.weightQ5_K[4096, 4096]
-
blk.26.attn_q.weightQ4_K[4096, 4096]
-
blk.26.attn_v.weightQ8_0[4096, 1024]
-
blk.26.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.26.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.26.ffn_gate_inp.weightF32[4096, 8]
-
blk.26.ffn_norm.weightF32[4096]
-
blk.26.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.27.attn_k.weightQ8_0[4096, 1024]
-
blk.27.attn_norm.weightF32[4096]
-
blk.27.attn_output.weightQ5_K[4096, 4096]
-
blk.27.attn_q.weightQ4_K[4096, 4096]
-
blk.27.attn_v.weightQ8_0[4096, 1024]
-
blk.27.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.27.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.27.ffn_gate_inp.weightF32[4096, 8]
-
blk.27.ffn_norm.weightF32[4096]
-
blk.27.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.28.attn_k.weightQ8_0[4096, 1024]
-
blk.28.attn_norm.weightF32[4096]
-
blk.28.attn_output.weightQ5_K[4096, 4096]
-
blk.28.attn_q.weightQ4_K[4096, 4096]
-
blk.28.attn_v.weightQ8_0[4096, 1024]
-
blk.28.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.28.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.28.ffn_gate_inp.weightF32[4096, 8]
-
blk.28.ffn_norm.weightF32[4096]
-
blk.28.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.29.attn_k.weightQ8_0[4096, 1024]
-
blk.29.attn_norm.weightF32[4096]
-
blk.29.attn_output.weightQ5_K[4096, 4096]
-
blk.29.attn_q.weightQ4_K[4096, 4096]
-
blk.29.attn_v.weightQ8_0[4096, 1024]
-
blk.29.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.29.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.29.ffn_gate_inp.weightF32[4096, 8]
-
blk.29.ffn_norm.weightF32[4096]
-
blk.29.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.30.attn_k.weightQ8_0[4096, 1024]
-
blk.30.attn_norm.weightF32[4096]
-
blk.30.attn_output.weightQ5_K[4096, 4096]
-
blk.30.attn_q.weightQ4_K[4096, 4096]
-
blk.30.attn_v.weightQ8_0[4096, 1024]
-
blk.30.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.30.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.30.ffn_gate_inp.weightF32[4096, 8]
-
blk.30.ffn_norm.weightF32[4096]
-
blk.30.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
blk.31.attn_k.weightQ8_0[4096, 1024]
-
blk.31.attn_norm.weightF32[4096]
-
blk.31.attn_output.weightQ5_K[4096, 4096]
-
blk.31.attn_q.weightQ4_K[4096, 4096]
-
blk.31.attn_v.weightQ8_0[4096, 1024]
-
blk.31.ffn_down_exps.weightQ4_K[14336, 4096, 8]
-
blk.31.ffn_gate_exps.weightQ4_K[4096, 14336, 8]
-
blk.31.ffn_gate_inp.weightF32[4096, 8]
-
blk.31.ffn_norm.weightF32[4096]
-
blk.31.ffn_up_exps.weightQ4_K[4096, 14336, 8]
-
output.weightQ6_K[4096, 32002]
-
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