From c704c778f6841bb56fa5200c1ee3dff4cea1aa5d Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Tue, 2 Apr 2024 21:35:13 +0300 Subject: [PATCH] convert : fix grok tensor names --- convert-hf-to-gguf.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/convert-hf-to-gguf.py b/convert-hf-to-gguf.py index f45b145a4..afa034a86 100755 --- a/convert-hf-to-gguf.py +++ b/convert-hf-to-gguf.py @@ -1351,7 +1351,7 @@ class GrokModel(Model): for wid in ["linear", "linear_1", "linear_v"]: full = True for xid in range(n_experts): - ename = f"transformer.decoder_layer.{bid}.moe.{xid}.{wid}" + ename = f"transformer.decoder_layer.{bid}.moe.{xid}.{wid}.weight" if ename not in experts: full = False break @@ -1360,7 +1360,7 @@ class GrokModel(Model): datas = [] for xid in range(n_experts): - ename = f"transformer.decoder_layer.{bid}.moe.{xid}.{wid}" + ename = f"transformer.decoder_layer.{bid}.moe.{xid}.{wid}.weight" datas.append(experts[ename]) del experts[ename]