Skip to content

Commit cf9ccef

Browse files
authored
fix variable names to align with the new transformer version (#1792)
SUMMARY: Update past_key_value to past_key_values. TEST PLAN: This should resolve the first three kv cache failures. I'm looking into the last one related to meta device. Signed-off-by: shanjiaz <[email protected]>
1 parent 4115d1f commit cf9ccef

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

src/llmcompressor/modifiers/quantization/calibration.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -247,7 +247,7 @@ def calibrate_kv_cache_input_hook(
247247
kv_cache to singleton QuantizedKVParameterCache.
248248
"""
249249
kv_cache = getattr(module, "kv_cache")
250-
kwargs["past_key_value"] = kv_cache
250+
kwargs["past_key_values"] = kv_cache
251251
kwargs["use_cache"] = False
252252
return args, kwargs
253253

0 commit comments

Comments
 (0)