[LLM] Fix llm models extension issue

[changwangss]

Type of Change

Fix llm models extension test issue

1. phi transformers version.
2. add requirements for baichuan/qwen/codegen.

Support codegen sq workflow, due to codegen use the remote tokenizer, so add trust_remote_code to get the correct tokenizer. transformers version requests 4.35.2.
fp32 with limit 20

numactl -m 0 -C 0-55 python run_generation.py     --model Salesforce/codegen25-7b-multi    --accuracy     --batch_size 20     --n_samples 20     --allow_code_execution     --temperature 0.2     --do_sample     --tasks "humaneval" --trust_remote_code True

Evaluating generations...
{
  "humaneval": {
    "pass@1": 0.5725,
    "pass@10": 0.649981867977224
  },
  "config": {
    "model": "llama",
    "temperature": 0.2,
    "n_samples": 20
  }
}
{'humaneval': {'pass@1': 0.5725, 'pass@10': 0.649981867977224}, 'config': {'model': 'llama', 'temperature': 0.2, 'n_samples': 20}}

int8 sq with alpha 0,5 calib_iters 5 limit 20

python run_generation.py     --model Salesforce/codegen25-7b-multi        --alpha 0.5       --accuracy     --batch_size 20     --n_samples 20     --allow_code_execution     --temperature 0.2     --do_sample     --tasks "humaneval" --limit 20 --trust_remote_code True --sq --int8 --calib_iters 5

Evaluating generations...
{
  "humaneval": {
    "pass@1": 0.3575,
    "pass@10": 0.4497261252679209
  },
  "config": {
    "model": "llama",
    "temperature": 0.2,
    "n_samples": 20
  }
}
{'humaneval': {'pass@1': 0.3575, 'pass@10': 0.4497261252679209}, 'config': {'model': 'llama', 'temperature': 0.2, 'n_samples': 20}}

Description

detail description
JIRA ticket: xxx

Expected Behavior & Potential Risk

the expected behavior that triggered by this PR

How has this PR been tested?

how to reproduce the test (including hardware information)

Dependency Change?

any library dependency introduced or removed

[VincyZhang]

https://inteltf-jenk.sh.intel.com/view/ITREX-release/job/ITREX-1.3-release-test/44/

[changwangss]

https://inteltf-jenk.sh.intel.com/view/ITREX-release/job/ITREX-1.3-release-test/47/

[VincyZhang]

https://inteltf-jenk.sh.intel.com/job/ITREX-1.3-release-test/49/

[changwangss]

local test qwen

python -u ./run_generation.py --model Qwen/Qwen-7B --alpha 0.9 --output_dir saved_results_qwen --trust_remote_code True --int8 --benchmark --batch_size 1  --_commit_hash  f7bc352f27bb1c02ee371a4576942a7d96c8bb97 --calib_iter 10  --sq

[VincyZhang]

CI all pass, ready for merge.