51 Pulls Updated 2 weeks ago
13e309bd56ec · 86MB
-
general.architecturellama
-
general.file_typeQ4_K_M
-
llama.attention.head_count12
-
llama.attention.head_count_kv12
-
llama.attention.layer_norm_rms_epsilon1e-05
-
llama.block_count12
-
llama.context_length2048
-
llama.embedding_length768
-
llama.feed_forward_length2048
-
llama.rope.dimension_count64
-
llama.rope.freq_base10000
-
llama.vocab_size32000
-
tokenizer.ggml.add_bos_tokentrue
-
tokenizer.ggml.add_eos_tokenfalse
-
tokenizer.ggml.bos_token_id1
-
tokenizer.ggml.eos_token_id2
-
tokenizer.ggml.modelllama
-
tokenizer.ggml.predefault
-
tokenizer.ggml.scores[0 0 0 0 0 ...]
-
tokenizer.ggml.token_type[2 3 3 6 6 ...]
-
tokenizer.ggml.tokens[<unk> <s> </s> <0x00> <0x01> ...]
-
tokenizer.ggml.unknown_token_id0
-
NameTypeShape
-
token_embd.weightQ4_K[768 32000]
-
blk.0.attn_norm.weightF32[768]
-
blk.0.ffn_down.weightQ6_K[2048 768]
-
blk.0.ffn_gate.weightQ4_K[768 2048]
-
blk.0.ffn_up.weightQ4_K[768 2048]
-
blk.0.ffn_norm.weightF32[768]
-
blk.0.attn_k.weightQ4_K[768 768]
-
blk.0.attn_output.weightQ4_K[768 768]
-
blk.0.attn_q.weightQ4_K[768 768]
-
blk.0.attn_v.weightQ6_K[768 768]
-
blk.1.attn_norm.weightF32[768]
-
blk.1.ffn_down.weightQ4_K[2048 768]
-
blk.1.ffn_gate.weightQ4_K[768 2048]
-
blk.1.ffn_up.weightQ4_K[768 2048]
-
blk.1.ffn_norm.weightF32[768]
-
blk.1.attn_k.weightQ4_K[768 768]
-
blk.1.attn_output.weightQ4_K[768 768]
-
blk.1.attn_q.weightQ4_K[768 768]
-
blk.1.attn_v.weightQ4_K[768 768]
-
blk.2.attn_norm.weightF32[768]
-
blk.2.ffn_down.weightQ4_K[2048 768]
-
blk.2.ffn_gate.weightQ4_K[768 2048]
-
blk.2.ffn_up.weightQ4_K[768 2048]
-
blk.2.ffn_norm.weightF32[768]
-
blk.2.attn_k.weightQ4_K[768 768]
-
blk.2.attn_output.weightQ4_K[768 768]
-
blk.2.attn_q.weightQ4_K[768 768]
-
blk.2.attn_v.weightQ4_K[768 768]
-
blk.3.attn_norm.weightF32[768]
-
blk.3.ffn_down.weightQ4_K[2048 768]
-
blk.3.ffn_gate.weightQ4_K[768 2048]
-
blk.3.ffn_up.weightQ4_K[768 2048]
-
blk.3.ffn_norm.weightF32[768]
-
blk.3.attn_k.weightQ4_K[768 768]
-
blk.3.attn_output.weightQ4_K[768 768]
-
blk.3.attn_q.weightQ4_K[768 768]
-
blk.3.attn_v.weightQ4_K[768 768]
-
blk.4.attn_norm.weightF32[768]
-
blk.4.ffn_down.weightQ6_K[2048 768]
-
blk.4.ffn_gate.weightQ4_K[768 2048]
-
blk.4.ffn_up.weightQ4_K[768 2048]
-
blk.4.ffn_norm.weightF32[768]
-
blk.4.attn_k.weightQ4_K[768 768]
-
blk.4.attn_output.weightQ4_K[768 768]
-
blk.4.attn_q.weightQ4_K[768 768]
-
blk.4.attn_v.weightQ6_K[768 768]
-
blk.5.attn_norm.weightF32[768]
-
blk.5.ffn_down.weightQ4_K[2048 768]
-
blk.5.ffn_gate.weightQ4_K[768 2048]
-
blk.5.ffn_up.weightQ4_K[768 2048]
-
blk.5.ffn_norm.weightF32[768]
-
blk.5.attn_k.weightQ4_K[768 768]
-
blk.5.attn_output.weightQ4_K[768 768]
-
blk.5.attn_q.weightQ4_K[768 768]
-
blk.5.attn_v.weightQ4_K[768 768]
-
blk.6.attn_norm.weightF32[768]
-
blk.6.ffn_down.weightQ4_K[2048 768]
-
blk.6.ffn_gate.weightQ4_K[768 2048]
-
blk.6.ffn_up.weightQ4_K[768 2048]
-
blk.6.ffn_norm.weightF32[768]
-
blk.6.attn_k.weightQ4_K[768 768]
-
blk.6.attn_output.weightQ4_K[768 768]
-
blk.6.attn_q.weightQ4_K[768 768]
-
blk.6.attn_v.weightQ4_K[768 768]
-
blk.7.attn_norm.weightF32[768]
-
blk.7.ffn_down.weightQ6_K[2048 768]
-
blk.7.ffn_gate.weightQ4_K[768 2048]
-
blk.7.ffn_up.weightQ4_K[768 2048]
-
blk.7.ffn_norm.weightF32[768]
-
blk.7.attn_k.weightQ4_K[768 768]
-
blk.7.attn_output.weightQ4_K[768 768]
-
blk.7.attn_q.weightQ4_K[768 768]
-
blk.7.attn_v.weightQ6_K[768 768]
-
blk.8.attn_norm.weightF32[768]
-
blk.8.ffn_down.weightQ6_K[2048 768]
-
blk.8.ffn_gate.weightQ4_K[768 2048]
-
blk.8.ffn_up.weightQ4_K[768 2048]
-
blk.8.ffn_norm.weightF32[768]
-
blk.8.attn_k.weightQ4_K[768 768]
-
blk.8.attn_output.weightQ4_K[768 768]
-
blk.8.attn_q.weightQ4_K[768 768]
-
blk.8.attn_v.weightQ6_K[768 768]
-
blk.9.attn_norm.weightF32[768]
-
blk.9.ffn_down.weightQ6_K[2048 768]
-
blk.9.ffn_gate.weightQ4_K[768 2048]
-
blk.9.ffn_up.weightQ4_K[768 2048]
-
blk.9.ffn_norm.weightF32[768]
-
blk.9.attn_k.weightQ4_K[768 768]
-
blk.9.attn_output.weightQ4_K[768 768]
-
blk.9.attn_q.weightQ4_K[768 768]
-
blk.9.attn_v.weightQ6_K[768 768]
-
blk.10.attn_norm.weightF32[768]
-
blk.10.ffn_down.weightQ4_K[2048 768]
-
blk.10.ffn_gate.weightQ4_K[768 2048]
-
blk.10.ffn_up.weightQ4_K[768 2048]
-
blk.10.ffn_norm.weightF32[768]
-
blk.10.attn_k.weightQ4_K[768 768]
-
blk.10.attn_output.weightQ4_K[768 768]
-
blk.10.attn_q.weightQ4_K[768 768]
-
blk.10.attn_v.weightQ4_K[768 768]
-
blk.11.attn_norm.weightF32[768]
-
blk.11.ffn_down.weightQ6_K[2048 768]
-
blk.11.ffn_gate.weightQ4_K[768 2048]
-
blk.11.ffn_up.weightQ4_K[768 2048]
-
blk.11.ffn_norm.weightF32[768]
-
blk.11.attn_k.weightQ4_K[768 768]
-
blk.11.attn_output.weightQ4_K[768 768]
-
blk.11.attn_q.weightQ4_K[768 768]
-
blk.11.attn_v.weightQ6_K[768 768]
-
output.weightQ6_K[768 32000]
-
output_norm.weightF32[768]
Metadata
Tensors
blk.0
blk.1
blk.2
blk.3
blk.4
blk.5
blk.6
blk.7
blk.8
blk.9
blk.10
blk.11