[Bug fix][Core] assert num_new_tokens == 1 fails when SamplingParams.n is not 1 and max_tokens is large & Add tests for preemption (#4451)

2024-05-02 11:24:13 +09:00
parent b8afa8b95a
commit 0d62fe58db
6 changed files with 172 additions and 13 deletions
--- a/tests/basic_correctness/test_chunked_prefill.py
+++ b/tests/basic_correctness/test_chunked_prefill.py
@@ -55,7 +55,6 @@ def test_models(
    )
    vllm_outputs = vllm_model.generate_greedy(example_prompts, max_tokens)
    del vllm_model
-    print(vllm_outputs[0])

    for i in range(len(example_prompts)):
        hf_output_ids, hf_output_str = hf_outputs[i]