@Wovchena can you please guide me that what possibly gone wrong for cpp-greedy_causal_lm-Chatglm3-6b and cpp-prompt_lookup_decoding_lm-ubuntu although both greedy_causal_lm and prompt_lookup_decoding_lm works for chatglm3-6b in my local machine

@Wovchena
I use the following command but doesn't seems like it is working:

git pull --rebase origin verify-chatglm-6b
git add resolved-file
git rebase --continue
git push origin verify-chatglm-6b --force

can you suggest me anything which I may have done wrong or missed.
Additionally my forked repo is up to date with openvinotoolkit:master
please help me in this matter!

You can checkout thirparty/openvino_tokenizers from master and push to you branch to align the tokenizers with master:

git checkout master thirdparty/openvino_tokenizers/
git add thirdparty/openvino_tokenizers/
git commit -m tokenizers
git push

Apparently that was wrong master. Try git checkout a0268cd5c5fe71ccbc4dc773b502075867c859fe thirdparty/openvino_tokenizers instead of git checkout master thirdparty/openvino_tokenizers/

@Wovchena
why prompt_lookup and greedy is are generating different output for chatglm3?here

- name: run and compare
        run: |
          source ./ov/setupvars.sh

          echo 'Code:```python
          def add(a, b):
              return a + b
          ```
          Question: Can you please add 2 and 3
          A:' > ./prompt.txt

          ./build/samples/cpp/prompt_lookup_decoding_lm/prompt_lookup_decoding_lm ./chatglm3-6b/ "$(<prompt.txt)" > predictions_prompt_lookup.txt
          ./build/samples/cpp/greedy_causal_lm/greedy_causal_lm ./chatglm3-6b/ "$(<prompt.txt)" > predictions_greedy.txt
          diff predictions_prompt_lookup.txt predictions_greedy.txt
          python -c "
          with open('predictions_greedy.txt', 'r') as f:
              predicted_greedy = f.readline()
          with open('predictions_prompt_lookup.txt', 'r') as f:
              predicted_prompt_lookup = f.readline()
          print(predicted_greedy)
          print(predicted_prompt_lookup)
          "
          echo "Prompt lookup" passed

because as per the error raised I think it is due to the different output generated by greedy and prompt lookup.
But it is probably generating same results for Qwen model since it is passing the tests.

I don't know. That is the point of this task to test if chatglm is fine. Apparently it's not. You can print the results to see them

@Wovchena
As I can see cpp-greedy_causal_lm-Chatglm3-6b is failing because the output generated by python sample and c++ sample are different and whereas cpp-prompt_lookup_decoding_lm-ubuntu is failing because prompt lookup and greedy generated text are different.
In both the cases I am using diff command to compare the generated text which may be causing assertion error , so will I only need to remove comparison part or remove whole cpp-greedy_causal_lm-Chatglm3-6b and cpp-prompt_lookup_decoding_lm-ubuntu` .
Also don't know why

1. Linux (Ubuntu 20.04, Python 3.9) / OpenVINO genai extension (cmake + wheel) (pull_request) Failing after 27m
2. Windows (VS 2019, Python 3.11) / OpenVINO genai extension (cmake + wheel) (pull_request)
3. macOS (12, Python 3.9) / OpenVINO genai extension (cmake + wheel) (pull_request)
   the above tests are failing?
   I also proceed as you mentioned earlier

git checkout a0268cd5c5fe71ccbc4dc773b502075867c859fe thirdparty/openvino_tokenizers
git add thirdparty/openvino_tokenizers/
git commit --amend
git push

You need to find out why C++ and Python greedy produce different outputs and fix it. Other model runs are aligned, that means the problem is not about the samples themselves.

This PR diff shows that you've modified thirdparty/openvino_tokenizers. Eliminate that diff. master's version of openvino_tokenizers should be kept. May be this is the reason for Linux (Ubuntu 20.04, Python 3.9) / OpenVINO genai extension (cmake + wheel) (pull_request) failure.