The model used is a quantized version of `Llama-3-Taiwan-70B-Instruct`. More details can be found on the website (https://huggingface.co/yentinglin/Llama-3-Taiwan-70B-Instruct)
110 Pulls Updated 4 months ago
7b0e90f9e6ab · 10.0GB
-
split.count2
-
split.no1
-
split.tensors.count723
-
NameTypeShape
-
blk.67.attn_k.weightQ6_K[8192, 1024]
-
blk.67.attn_output.weightQ6_K[8192, 8192]
-
blk.67.attn_q.weightQ6_K[8192, 8192]
-
blk.67.attn_v.weightQ6_K[8192, 1024]
-
blk.67.ffn_down.weightQ6_K[28672, 8192]
-
blk.67.ffn_gate.weightQ6_K[8192, 28672]
-
blk.67.ffn_norm.weightF32[8192]
-
blk.67.ffn_up.weightQ6_K[8192, 28672]
-
blk.68.attn_k.weightQ6_K[8192, 1024]
-
blk.68.attn_norm.weightF32[8192]
-
blk.68.attn_output.weightQ6_K[8192, 8192]
-
blk.68.attn_q.weightQ6_K[8192, 8192]
-
blk.68.attn_v.weightQ6_K[8192, 1024]
-
blk.68.ffn_down.weightQ6_K[28672, 8192]
-
blk.68.ffn_gate.weightQ6_K[8192, 28672]
-
blk.68.ffn_norm.weightF32[8192]
-
blk.68.ffn_up.weightQ6_K[8192, 28672]
-
blk.69.attn_k.weightQ6_K[8192, 1024]
-
blk.69.attn_norm.weightF32[8192]
-
blk.69.attn_output.weightQ6_K[8192, 8192]
-
blk.69.attn_q.weightQ6_K[8192, 8192]
-
blk.69.attn_v.weightQ6_K[8192, 1024]
-
blk.69.ffn_down.weightQ6_K[28672, 8192]
-
blk.69.ffn_gate.weightQ6_K[8192, 28672]
-
blk.69.ffn_norm.weightF32[8192]
-
blk.69.ffn_up.weightQ6_K[8192, 28672]
-
blk.70.attn_k.weightQ6_K[8192, 1024]
-
blk.70.attn_norm.weightF32[8192]
-
blk.70.attn_output.weightQ6_K[8192, 8192]
-
blk.70.attn_q.weightQ6_K[8192, 8192]
-
blk.70.attn_v.weightQ6_K[8192, 1024]
-
blk.70.ffn_down.weightQ6_K[28672, 8192]
-
blk.70.ffn_gate.weightQ6_K[8192, 28672]
-
blk.70.ffn_norm.weightF32[8192]
-
blk.70.ffn_up.weightQ6_K[8192, 28672]
-
blk.71.attn_k.weightQ6_K[8192, 1024]
-
blk.71.attn_norm.weightF32[8192]
-
blk.71.attn_output.weightQ6_K[8192, 8192]
-
blk.71.attn_q.weightQ6_K[8192, 8192]
-
blk.71.attn_v.weightQ6_K[8192, 1024]
-
blk.71.ffn_down.weightQ6_K[28672, 8192]
-
blk.71.ffn_gate.weightQ6_K[8192, 28672]
-
blk.71.ffn_norm.weightF32[8192]
-
blk.71.ffn_up.weightQ6_K[8192, 28672]
-
blk.72.attn_k.weightQ6_K[8192, 1024]
-
blk.72.attn_norm.weightF32[8192]
-
blk.72.attn_output.weightQ6_K[8192, 8192]
-
blk.72.attn_q.weightQ6_K[8192, 8192]
-
blk.72.attn_v.weightQ6_K[8192, 1024]
-
blk.72.ffn_down.weightQ6_K[28672, 8192]
-
blk.72.ffn_gate.weightQ6_K[8192, 28672]
-
blk.72.ffn_norm.weightF32[8192]
-
blk.72.ffn_up.weightQ6_K[8192, 28672]
-
blk.73.attn_k.weightQ6_K[8192, 1024]
-
blk.73.attn_norm.weightF32[8192]
-
blk.73.attn_output.weightQ6_K[8192, 8192]
-
blk.73.attn_q.weightQ6_K[8192, 8192]
-
blk.73.attn_v.weightQ6_K[8192, 1024]
-
blk.73.ffn_down.weightQ6_K[28672, 8192]
-
blk.73.ffn_gate.weightQ6_K[8192, 28672]
-
blk.73.ffn_norm.weightF32[8192]
-
blk.73.ffn_up.weightQ6_K[8192, 28672]
-
blk.74.attn_k.weightQ6_K[8192, 1024]
-
blk.74.attn_norm.weightF32[8192]
-
blk.74.attn_output.weightQ6_K[8192, 8192]
-
blk.74.attn_q.weightQ6_K[8192, 8192]
-
blk.74.attn_v.weightQ6_K[8192, 1024]
-
blk.74.ffn_down.weightQ6_K[28672, 8192]
-
blk.74.ffn_gate.weightQ6_K[8192, 28672]
-
blk.74.ffn_norm.weightF32[8192]
-
blk.74.ffn_up.weightQ6_K[8192, 28672]
-
blk.75.attn_k.weightQ6_K[8192, 1024]
-
blk.75.attn_norm.weightF32[8192]
-
blk.75.attn_output.weightQ6_K[8192, 8192]
-
blk.75.attn_q.weightQ6_K[8192, 8192]
-
blk.75.attn_v.weightQ6_K[8192, 1024]
-
blk.75.ffn_down.weightQ6_K[28672, 8192]
-
blk.75.ffn_gate.weightQ6_K[8192, 28672]
-
blk.75.ffn_norm.weightF32[8192]
-
blk.75.ffn_up.weightQ6_K[8192, 28672]
-
blk.76.attn_k.weightQ6_K[8192, 1024]
-
blk.76.attn_norm.weightF32[8192]
-
blk.76.attn_output.weightQ6_K[8192, 8192]
-
blk.76.attn_q.weightQ6_K[8192, 8192]
-
blk.76.attn_v.weightQ6_K[8192, 1024]
-
blk.76.ffn_down.weightQ6_K[28672, 8192]
-
blk.76.ffn_gate.weightQ6_K[8192, 28672]
-
blk.76.ffn_norm.weightF32[8192]
-
blk.76.ffn_up.weightQ6_K[8192, 28672]
-
blk.77.attn_k.weightQ6_K[8192, 1024]
-
blk.77.attn_norm.weightF32[8192]
-
blk.77.attn_output.weightQ6_K[8192, 8192]
-
blk.77.attn_q.weightQ6_K[8192, 8192]
-
blk.77.attn_v.weightQ6_K[8192, 1024]
-
blk.77.ffn_down.weightQ6_K[28672, 8192]
-
blk.77.ffn_gate.weightQ6_K[8192, 28672]
-
blk.77.ffn_norm.weightF32[8192]
-
blk.77.ffn_up.weightQ6_K[8192, 28672]
-
blk.78.attn_k.weightQ6_K[8192, 1024]
-
blk.78.attn_norm.weightF32[8192]
-
blk.78.attn_output.weightQ6_K[8192, 8192]
-
blk.78.attn_q.weightQ6_K[8192, 8192]
-
blk.78.attn_v.weightQ6_K[8192, 1024]
-
blk.78.ffn_down.weightQ6_K[28672, 8192]
-
blk.78.ffn_gate.weightQ6_K[8192, 28672]
-
blk.78.ffn_norm.weightF32[8192]
-
blk.78.ffn_up.weightQ6_K[8192, 28672]
-
blk.79.attn_k.weightQ6_K[8192, 1024]
-
blk.79.attn_norm.weightF32[8192]
-
blk.79.attn_output.weightQ6_K[8192, 8192]
-
blk.79.attn_q.weightQ6_K[8192, 8192]
-
blk.79.attn_v.weightQ6_K[8192, 1024]
-
blk.79.ffn_down.weightQ6_K[28672, 8192]
-
blk.79.ffn_gate.weightQ6_K[8192, 28672]
-
blk.79.ffn_norm.weightF32[8192]
-
blk.79.ffn_up.weightQ6_K[8192, 28672]
-
output.weightQ6_K[8192, 128256]
-
output_norm.weightF32[8192]
Metadata
Tensor
blk.67
blk.68
blk.69
blk.70
blk.71
blk.72
blk.73
blk.74
blk.75
blk.76
blk.77
blk.78
blk.79