biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Core][Model] Terratorch backend integration (#23513)

Browse Source 
Signed-off-by: Michele Gazzetti <michele.gazzetti1@ibm.com>
Signed-off-by: Christian Pinto <christian.pinto@ibm.com>
Co-authored-by: Christian Pinto <christian.pinto@ibm.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
This commit is contained in:
mgazz
2025-09-04 08:22:41 +01:00
committed by GitHub
parent e7fc70016f
commit 51d5e9be7d
23 changed files with 305 additions and 208 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
tests/distributed/test_pipeline_parallel.py
View File

@@ -298,6 +298,8 @@ def _compare_tp(
tokenizer_mode = model_info.tokenizer_mode
hf_overrides = model_info.hf_overrides
hf_config = get_config(model_id, trust_remote_code)
skip_tokenizer_init = model_info.skip_tokenizer_init
max_num_seqs = model_info.max_num_seqs
dtype = "float16"
if hf_config.model_type in _FLOAT16_NOT_SUPPORTED_MODELS:
@@ -351,6 +353,10 @@ def _compare_tp(
common_args.extend(["--load-format", load_format])
if hf_overrides:
common_args.extend(["--hf-overrides", json.dumps(hf_overrides)])
if skip_tokenizer_init:
common_args.append("--skip-tokenizer-init")
if max_num_seqs:
common_args.extend(["--max-num-seqs", f"{max_num_seqs}"])
specific_case = tp_size == 2 and pp_size == 2 and chunked_prefill
testing_ray_compiled_graph = False