Dolphin 3.0 Llama 3.1 8B 🐬 is the next generation of the Dolphin series of instruct-tuned models designed to be the ultimate general purpose local model, enabling coding, math, agentic, function calling, and general use cases.
8b
353.5K Pulls Updated 2 months ago
1eee69535308 · 4.9GB
-
general.architecturellamallama
-
general.base_model.0.nameLlama 3.1 8BLlama 3.1 8B
-
general.base_model.0.organizationMeta LlamaMeta Llama
-
general.base_model.0.repo_urlhttps://huggingface.co/meta-llama/Llama-3.1-8Bhttps://huggingface.co/meta-llama/Llama-3.1-8B
-
general.base_model.count11
-
general.basenameDolphin-3.0-Llama-3.1Dolphin-3.0-Llama-3.1
-
general.dataset.0.nameOpc Sft Stage1Opc Sft Stage1
-
general.dataset.0.organizationOpenCoder LLMOpenCoder LLM
-
general.dataset.0.repo_urlhttps://huggingface.co/OpenCoder-LLM/opc-sft-stage1https://huggingface.co/OpenCoder-LLM/opc-sft-stage1
-
general.dataset.1.nameOpc Sft Stage2Opc Sft Stage2
-
general.dataset.1.organizationOpenCoder LLMOpenCoder LLM
-
general.dataset.1.repo_urlhttps://huggingface.co/OpenCoder-LLM/opc-sft-stage2https://huggingface.co/OpenCoder-LLM/opc-sft-stage2
-
general.dataset.10.nameSamantha DataSamantha Data
-
general.dataset.10.organizationCognitivecomputationsCognitivecomputations
-
general.dataset.10.repo_urlhttps://huggingface.co/cognitivecomputations/samantha-datahttps://huggingface.co/cognitivecomputations/samantha-data
-
general.dataset.11.nameCodeFeedback Filtered InstructionCodeFeedback Filtered Instruction
-
general.dataset.11.organizationM A PM A P
-
general.dataset.11.repo_urlhttps://huggingface.co/m-a-p/CodeFeedback-Filtered-Instructionhttps://huggingface.co/m-a-p/CodeFeedback-Filtered-Instruction
-
general.dataset.12.nameCode FeedbackCode Feedback
-
general.dataset.12.organizationM A PM A P
-
general.dataset.12.repo_urlhttps://huggingface.co/m-a-p/Code-Feedbackhttps://huggingface.co/m-a-p/Code-Feedback
-
general.dataset.2.nameOrca Agentinstruct 1M v1Orca Agentinstruct 1M v1
-
general.dataset.2.organizationMicrosoftMicrosoft
-
general.dataset.2.repo_urlhttps://huggingface.co/microsoft/orca-agentinstruct-1M-v1https://huggingface.co/microsoft/orca-agentinstruct-1M-v1
-
general.dataset.2.versionv1v1
-
general.dataset.3.nameOrca Math Word Problems 200kOrca Math Word Problems 200k
-
general.dataset.3.organizationMicrosoftMicrosoft
-
general.dataset.3.repo_urlhttps://huggingface.co/microsoft/orca-math-word-problems-200khttps://huggingface.co/microsoft/orca-math-word-problems-200k
-
general.dataset.4.nameHermes Function Calling v1Hermes Function Calling v1
-
general.dataset.4.organizationNousResearchNousResearch
-
general.dataset.4.repo_urlhttps://huggingface.co/NousResearch/hermes-function-calling-v1https://huggingface.co/NousResearch/hermes-function-calling-v1
-
general.dataset.4.versionv1v1
-
general.dataset.5.nameNuminaMath CoTNuminaMath CoT
-
general.dataset.5.organizationAI MOAI MO
-
general.dataset.5.repo_urlhttps://huggingface.co/AI-MO/NuminaMath-CoThttps://huggingface.co/AI-MO/NuminaMath-CoT
-
general.dataset.6.nameNuminaMath TIRNuminaMath TIR
-
general.dataset.6.organizationAI MOAI MO
-
general.dataset.6.repo_urlhttps://huggingface.co/AI-MO/NuminaMath-TIRhttps://huggingface.co/AI-MO/NuminaMath-TIR
-
general.dataset.7.nameTulu 3 Sft MixtureTulu 3 Sft Mixture
-
general.dataset.7.organizationAllenaiAllenai
-
general.dataset.7.repo_urlhttps://huggingface.co/allenai/tulu-3-sft-mixturehttps://huggingface.co/allenai/tulu-3-sft-mixture
-
general.dataset.8.nameDolphin CoderDolphin Coder
-
general.dataset.8.organizationCognitivecomputationsCognitivecomputations
-
general.dataset.8.repo_urlhttps://huggingface.co/cognitivecomputations/dolphin-coderhttps://huggingface.co/cognitivecomputations/dolphin-coder
-
general.dataset.9.nameSmoltalkSmoltalk
-
general.dataset.9.organizationHuggingFaceTBHuggingFaceTB
-
general.dataset.9.repo_urlhttps://huggingface.co/HuggingFaceTB/smoltalkhttps://huggingface.co/HuggingFaceTB/smoltalk
-
general.dataset.count1313
-
general.file_type1515
-
general.languages[en][en]
-
general.licensellama3.1llama3.1
-
general.nameDolphin 3.0 Llama 3.1 8BDolphin 3.0 Llama 3.1 8B
-
general.organizationCognitivecomputationsCognitivecomputations
-
general.quantization_version22
-
general.size_label8B8B
-
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_length131072131072
-
llama.embedding_length40964096
-
llama.feed_forward_length1433614336
-
llama.rope.dimension_count128128
-
llama.rope.freq_base500000500000
-
llama.vocab_size128258128258
-
tokenizer.ggml.bos_token_id128000128000
-
tokenizer.ggml.eos_token_id128256128256
-
tokenizer.ggml.merges[Ġ Ġ, Ġ ĠĠĠ, ĠĠ ĠĠ, ĠĠĠ Ġ, i n, ...][Ġ Ġ, Ġ ĠĠĠ, ĠĠ ĠĠ, ĠĠĠ Ġ, i n, ...]
-
tokenizer.ggml.modelgpt2gpt2
-
tokenizer.ggml.padding_token_id128001128001
-
tokenizer.ggml.prellama-bpellama-bpe
-
tokenizer.ggml.token_type[1, 1, 1, 1, 1, ...][1, 1, 1, 1, 1, ...]
-
tokenizer.ggml.tokens[!, ", #, $, %, ...][!, ", #, $, %, ...]
-
token_embd.weightQ4_K[4096, 128258]
-
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[14336, 4096]
-
blk.0.ffn_gate.weightQ4_K[4096, 14336]
-
blk.0.ffn_norm.weightF32[4096]
-
blk.0.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.1.ffn_gate.weightQ4_K[4096, 14336]
-
blk.1.ffn_norm.weightF32[4096]
-
blk.1.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.2.ffn_gate.weightQ4_K[4096, 14336]
-
blk.2.ffn_norm.weightF32[4096]
-
blk.2.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.3.ffn_gate.weightQ4_K[4096, 14336]
-
blk.3.ffn_norm.weightF32[4096]
-
blk.3.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.4.ffn_gate.weightQ4_K[4096, 14336]
-
blk.4.ffn_norm.weightF32[4096]
-
blk.4.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.5.ffn_gate.weightQ4_K[4096, 14336]
-
blk.5.ffn_norm.weightF32[4096]
-
blk.5.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.6.ffn_gate.weightQ4_K[4096, 14336]
-
blk.6.ffn_norm.weightF32[4096]
-
blk.6.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.7.ffn_gate.weightQ4_K[4096, 14336]
-
blk.7.ffn_norm.weightF32[4096]
-
blk.7.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.8.ffn_gate.weightQ4_K[4096, 14336]
-
blk.8.ffn_norm.weightF32[4096]
-
blk.8.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ4_K[14336, 4096]
-
blk.9.ffn_gate.weightQ4_K[4096, 14336]
-
blk.9.ffn_norm.weightF32[4096]
-
blk.9.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ6_K[14336, 4096]
-
blk.10.ffn_gate.weightQ4_K[4096, 14336]
-
blk.10.ffn_norm.weightF32[4096]
-
blk.10.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ4_K[14336, 4096]
-
blk.11.ffn_gate.weightQ4_K[4096, 14336]
-
blk.11.ffn_norm.weightF32[4096]
-
blk.11.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.12.ffn_gate.weightQ4_K[4096, 14336]
-
blk.12.ffn_norm.weightF32[4096]
-
blk.12.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ6_K[14336, 4096]
-
blk.13.ffn_gate.weightQ4_K[4096, 14336]
-
blk.13.ffn_norm.weightF32[4096]
-
blk.13.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ4_K[14336, 4096]
-
blk.14.ffn_gate.weightQ4_K[4096, 14336]
-
blk.14.ffn_norm.weightF32[4096]
-
blk.14.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.15.ffn_gate.weightQ4_K[4096, 14336]
-
blk.15.ffn_norm.weightF32[4096]
-
blk.15.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ6_K[14336, 4096]
-
blk.16.ffn_gate.weightQ4_K[4096, 14336]
-
blk.16.ffn_norm.weightF32[4096]
-
blk.16.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ4_K[14336, 4096]
-
blk.17.ffn_gate.weightQ4_K[4096, 14336]
-
blk.17.ffn_norm.weightF32[4096]
-
blk.17.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.18.ffn_gate.weightQ4_K[4096, 14336]
-
blk.18.ffn_norm.weightF32[4096]
-
blk.18.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ6_K[14336, 4096]
-
blk.19.ffn_gate.weightQ4_K[4096, 14336]
-
blk.19.ffn_norm.weightF32[4096]
-
blk.19.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ4_K[14336, 4096]
-
blk.20.ffn_gate.weightQ4_K[4096, 14336]
-
blk.20.ffn_norm.weightF32[4096]
-
blk.20.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ6_K[4096, 1024]
-
blk.21.ffn_down.weightQ6_K[14336, 4096]
-
blk.21.ffn_gate.weightQ4_K[4096, 14336]
-
blk.21.ffn_norm.weightF32[4096]
-
blk.21.ffn_up.weightQ4_K[4096, 14336]
-
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.weightQ4_K[4096, 1024]
-
blk.22.ffn_down.weightQ4_K[14336, 4096]
-
blk.22.ffn_gate.weightQ4_K[4096, 14336]
-
blk.22.ffn_norm.weightF32[4096]
-
blk.22.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.23.ffn_gate.weightQ4_K[4096, 14336]
-
blk.23.ffn_norm.weightF32[4096]
-
blk.23.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.24.ffn_gate.weightQ4_K[4096, 14336]
-
blk.24.ffn_norm.weightF32[4096]
-
blk.24.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.25.ffn_gate.weightQ4_K[4096, 14336]
-
blk.25.ffn_norm.weightF32[4096]
-
blk.25.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.26.ffn_gate.weightQ4_K[4096, 14336]
-
blk.26.ffn_norm.weightF32[4096]
-
blk.26.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.27.ffn_gate.weightQ4_K[4096, 14336]
-
blk.27.ffn_norm.weightF32[4096]
-
blk.27.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.28.ffn_gate.weightQ4_K[4096, 14336]
-
blk.28.ffn_norm.weightF32[4096]
-
blk.28.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.29.ffn_gate.weightQ4_K[4096, 14336]
-
blk.29.ffn_norm.weightF32[4096]
-
blk.29.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.30.ffn_gate.weightQ4_K[4096, 14336]
-
blk.30.ffn_norm.weightF32[4096]
-
blk.30.ffn_up.weightQ4_K[4096, 14336]
-
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[14336, 4096]
-
blk.31.ffn_gate.weightQ4_K[4096, 14336]
-
blk.31.ffn_norm.weightF32[4096]
-
blk.31.ffn_up.weightQ4_K[4096, 14336]
-
output.weightQ6_K[4096, 128258]
-
rope_freqs.weightF32[64]
-
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