From 0635d0e24590d36bbfc76229e0882c49308d424c Mon Sep 17 00:00:00 2001
From: Nicolas Patry <patry.nicolas@protonmail.com>
Date: Tue, 25 Jul 2023 09:14:47 +0200
Subject: [PATCH] After rebase.

---
 server/text_generation_server/utils/weights.py | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/server/text_generation_server/utils/weights.py b/server/text_generation_server/utils/weights.py
index 3bfbf22c..0330402d 100644
--- a/server/text_generation_server/utils/weights.py
+++ b/server/text_generation_server/utils/weights.py
@@ -187,11 +187,8 @@ class Weights:
                     qzeros = self.get_sharded(f"{prefix}.qzeros", dim=0)
                     scales = self.get_sharded(f"{prefix}.scales", dim=0)
                 else:
-                    raise RuntimeError(
-                        "Using exllama GPTQ kernel with groupsize<1 is not supported"
-                    )
-                    # qzeros = self.get_tensor(f"{prefix}.qzeros")
-                    # scales = self.get_tensor(f"{prefix}.scales")
+                    qzeros = self.get_tensor(f"{prefix}.qzeros")
+                    scales = self.get_tensor(f"{prefix}.scales")
 
                 # For tp > 1, at this point we know we do not use act-order
                 if self.process_group.size() == 1: