From b742b5549ec2796be0d6f98f0dafd070cb093309 Mon Sep 17 00:00:00 2001
From: Christian Medina <37550954+cmedinasoriano@users.noreply.github.com>
Date: Thu, 2 Jul 2026 23:32:38 -0400
Subject: [PATCH] fix: alternate GPUs for each shard (0,1,0,1)

---
 quantize_streaming.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/quantize_streaming.py b/quantize_streaming.py
index 698faec..b6ad47f 100644
--- a/quantize_streaming.py
+++ b/quantize_streaming.py
@@ -96,7 +96,7 @@ def streaming_quantize(model_path: str, output_path: str):
             futures = []
             for j, shard_file in enumerate(batch):
                 idx = i + j
-                gpu_id = (j // shards_per_gpu) % num_gpus  # Assign to GPU
+                gpu_id = idx % num_gpus  # Alternate GPUs: 0,1,0,1...
                 futures.append(executor.submit(process_shard, idx, shard_file, gpu_id))
             for future in concurrent.futures.as_completed(futures):
                 future.result()