ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

LLM: add star_corder_15.5b model (#9772)

Browse source 
* LLM: add star_corder_15.5b model

* revert llm_performance_tests.yml
This commit is contained in:
WeiguangHan 2023-12-26 18:55:56 +08:00 committed by GitHub
parent 44b4a0c9c5
commit c05d7e1532
2 changed files with 7 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
python/llm/dev/benchmark/all-in-one/run.py
View file

@ -372,8 +372,12 @@ def run_transformer_int4_gpu(repo_id,
tokenizer = LlamaTokenizer.from_pretrained(model_path, trust_remote_code=True)
model = model.to('xpu')
else:
model = AutoModelForCausalLM.from_pretrained(model_path, optimize_model=True, load_in_low_bit=low_bit,
trust_remote_code=True, use_cache=True).eval()
if 'starcoder' in repo_id:
model = AutoModelForCausalLM.from_pretrained(model_path, optimize_model=True, load_in_low_bit=low_bit,
trust_remote_code=True, use_cache=True, torch_dtype=torch.bfloat16).eval()
else:
model = AutoModelForCausalLM.from_pretrained(model_path, optimize_model=True, load_in_low_bit=low_bit,
trust_remote_code=True, use_cache=True).eval()
tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
model = model.to('xpu')
if isinstance(model, GPTJForCausalLM):

2
python/llm/test/benchmark/arc-perf-test.yaml
View file

@ -5,7 +5,7 @@ repo_id:
- 'tiiuae/falcon-7b-instruct-with-patch'
- 'mosaicml/mpt-7b-chat'
- 'redpajama/gptneox-7b-redpajama-bf16'
# - 'bigcode/starcoder-15.5b'
- 'bigcode/starcoder-15.5b'
- 'databricks/dolly-v1-6b'
- 'databricks/dolly-v2-7b'
- 'databricks/dolly-v2-12b'