Adding Compute-Context-Length(CCL)

vjanfaza · vjanfaza · commit acf35442fa0f · 2025-10-22T17:32:04.000-07:00
Signed-off-by: Vahid Janfaza &lt;vjanfaza@qti.qualcomm.com&gt;
diff --git a/examples/compute_context_length.py b/examples/compute_context_length.py
@@ -41,9 +41,6 @@
 model = QEFFAutoModelForCausalLM.from_pretrained(
     model_name,
     continuous_batching=True,
-    comp_ctx_lengths_prefill=comp_ctx_lengths_prefill,
-    comp_ctx_lengths_decode=comp_ctx_lengths_decode,
-    ctx_len=ctx_len,
 )
 
 # model compilation for either continuous or static batching. For continuous batching full_batch_size is needed.

Original file line number	Diff line number	Diff line change
`@@ -41,9 +41,6 @@`
`41`	`41`	`model = QEFFAutoModelForCausalLM.from_pretrained(`
`42`	`42`	`model_name,`
`43`	`43`	`continuous_batching=True,`
`44`		`- comp_ctx_lengths_prefill=comp_ctx_lengths_prefill,`
`45`		`- comp_ctx_lengths_decode=comp_ctx_lengths_decode,`
`46`		`- ctx_len=ctx_len,`
`47`	`44`	`)`
`48`	`45`
`49`	`46`	`# model compilation for either continuous or static batching. For continuous batching full_batch_size is needed.`