missing argument ('repo_id') & dependency installation error

[pyongjoo]

• I git-cloned the current main.
• I am testing this repo: https://bolobao.ai/models/tscholak__SEP__1zha5ono.
• Environment: Mac Apple Silicon.
• Python 3.13.11

Missing Argument

Tested this model with the following script.

from bolo import pipeline

result = pipeline(
    model="tscholak/1zha5ono",
    device="cpu",
    text_input="How many singers do we have? | concert_singer | stadium : stadium_id, location, name, capacity, highest, lowest, average | singer : singer_id, name, country, song_name, song_release_year, age, is_male | concert : concert_id, concert_name, theme, stadium_id, year | singer_in_concert : concert_id, singer_id",
    num_beams=4,
    max_length=512,
)
print(result)

Expected Result: it runs.

Error:

(venv) May-05T10:48:51 yongjoo:workspace/Bolo/ (main) % python test.py 
Traceback (most recent call last):
  File "/Users/yongjoo/workspace/Bolo/test.py", line 3, in <module>
    result = pipeline(
        model="tscholak/1zha5ono",
    ...<3 lines>...
        max_length=512,
    )
TypeError: pipeline() missing 1 required positional argument: 'repo_id'

Claude could fix the error as follows:

from bolo import pipeline


def test_tscholak_1zha5ono():
    result = pipeline(
        "tscholak/1zha5ono",
        device="cpu",
        text_input=(
            "How many singers do we have? | concert_singer | "
            "stadium : stadium_id, location, name, capacity, highest, lowest, average | "
            "singer : singer_id, name, country, song_name, song_release_year, age, is_male | "
            "concert : concert_id, concert_name, theme, stadium_id, year | "
            "singer_in_concert : concert_id, singer_id"
        ),
        num_beams=4,
        max_length=512,
    )
    assert isinstance(result, dict)
    assert "generated_sql" in result
    sql = result["generated_sql"]
    assert isinstance(sql, str) and len(sql) > 0
    print(f"generated SQL: {sql}")

Dependency Error

Edit by Yunqi: I think this is expected since for reproducibility we just provide the dependencies for cu128. Then installing on any other platforms will fail, and I denoted this in the README.md file.

Steps:

1. I copied the requirements.txt from the model card (I wish there were a bolo command for installing them without creating requirements.txt)
2. I ran pip install -r requirements.txt --extra-index-url https://download.pytorch.org/whl/cu128
3. The following error happened.

(venv) May-05T10:53:08 yongjoo:workspace/Bolo/ (main) % pip install -r requirements.txt --extra-index-url https://download.pytorch.org/whl/cu128
Looking in indexes: https://pypi.org/simple, https://download.pytorch.org/whl/cu128
Collecting annotated-doc==0.0.4 (from -r requirements.txt (line 1))
  Using cached annotated_doc-0.0.4-py3-none-any.whl.metadata (6.6 kB)
Collecting anyio==4.13.0 (from -r requirements.txt (line 2))
  Using cached anyio-4.13.0-py3-none-any.whl.metadata (4.5 kB)
Collecting certifi==2022.12.7 (from -r requirements.txt (line 3))
  Downloading https://download.pytorch.org/whl/certifi-2022.12.7-py3-none-any.whl (155 kB)
Collecting click==8.3.2 (from -r requirements.txt (line 4))
  Using cached click-8.3.2-py3-none-any.whl.metadata (2.6 kB)
ERROR: Could not find a version that satisfies the requirement cuda-bindings==12.9.4 (from versions: none)

[pyongjoo]

To run cpu mode, I removed cuda dependencies and installed the remaining ones.

Now I see the following error. Cursor/Sonnet says it's because our template is overriding device config.

(venv) May-05T11:05:18 yongjoo:workspace/Bolo/ (main) % python test.py 
repo_id: tscholak/1zha5ono
num_params: 4
name       | type | default                                                                                                                                                                                                                                                                                                            | raw_default                                                                                                                                                                                                                                                                                                         
-----------+------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
device     | str  | cuda:0                                                                                                                                                                                                                                                                                                             | "cuda:0"                                                                                                                                                                                                                                                                                                            
text_input | str  | How many singers do we have? | concert_singer | stadium : stadium_id, location, name, capacity, highest, lowest, average | singer : singer_id, name, country, song_name, song_release_year, age, is_male | concert : concert_id, concert_name, theme, stadium_id, year | singer_in_concert : concert_id, singer_id | "How many singers do we have? | concert_singer | stadium : stadium_id, location, name, capacity, highest, lowest, average | singer : singer_id, name, country, song_name, song_release_year, age, is_male | concert : concert_id, concert_name, theme, stadium_id, year | singer_in_concert : concert_id, singer_id"
num_beams  | int  | 4                                                                                                                                                                                                                                                                                                                  | 4                                                                                                                                                                                                                                                                                                                   
max_length | int  | 512                                                                                                                                                                                                                                                                                                                | 512                                                                                                                                                                                                                                                                                                                 
Loading weights: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 284/284 [00:00<00:00, 37957.50it/s]
The tied weights mapping and config for this model specifies to tie shared.weight to lm_head.weight, but both are present in the checkpoints with different values, so we will NOT tie them. You should update the config with `tie_word_embeddings=False` to silence this warning.
Traceback (most recent call last):
  File "/Users/yongjoo/workspace/Bolo/test.py", line 3, in <module>
    result = pipeline(
        "tscholak/1zha5ono",
    ...<3 lines>...
        max_length=512,
    )
  File "/Users/yongjoo/workspace/Bolo/src/bolo/api.py", line 316, in pipeline
    return client.run_inference(code)
           ~~~~~~~~~~~~~~~~~~~~^^^^^^
  File "/Users/yongjoo/workspace/Bolo/src/bolo/api.py", line 266, in run_inference
    exec(code, ns, ns)
    ~~~~^^^^^^^^^^^^^^
  File "<string>", line 16, in <module>
  File "/Users/yongjoo/workspace/Bolo/venv/lib/python3.13/site-packages/transformers/modeling_utils.py", line 3574, in to
    return super().to(*args, **kwargs)
           ~~~~~~~~~~^^^^^^^^^^^^^^^^^
  File "/Users/yongjoo/workspace/Bolo/venv/lib/python3.13/site-packages/torch/nn/modules/module.py", line 1384, in to
    return self._apply(convert)
           ~~~~~~~~~~~^^^^^^^^^
  File "/Users/yongjoo/workspace/Bolo/venv/lib/python3.13/site-packages/torch/nn/modules/module.py", line 934, in _apply
    module._apply(fn)
    ~~~~~~~~~~~~~^^^^
  File "/Users/yongjoo/workspace/Bolo/venv/lib/python3.13/site-packages/torch/nn/modules/module.py", line 965, in _apply
    param_applied = fn(param)
  File "/Users/yongjoo/workspace/Bolo/venv/lib/python3.13/site-packages/torch/nn/modules/module.py", line 1370, in convert
    return t.to(
           ~~~~^
        device,
        ^^^^^^^
        dtype if t.is_floating_point() or t.is_complex() else None,
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        non_blocking,
        ^^^^^^^^^^^^^
    )
    ^
  File "/Users/yongjoo/workspace/Bolo/venv/lib/python3.13/site-packages/torch/cuda/__init__.py", line 471, in _lazy_init
    raise AssertionError("Torch not compiled with CUDA enabled")
AssertionError: Torch not compiled with CUDA enabled

The following patch could fix the error (at least for the model):

--- a/template.j2
+++ b/template.j2
@@ -1,6 +1,6 @@
-{% set device = "cuda:0" %}
-{% set repo_id = "tscholak/1zha5ono" %}
-{% set text_input = "How many singers do we have? | concert_singer | stadium : stadium_id, location, name, capacity, highest, lowest, average | singer : singer_id, name, country, song_name, song_release_year, age, is_male | concert : concert_id, concert_name, theme, stadium_id, year | singer_in_concert : concert_id, singer_id" %}
-{% set num_beams = 4 %}
-{% set max_length = 512 %}
+{% set device = device | default("cuda:0") %}
+{% set repo_id = repo_id | default("tscholak/1zha5ono") %}
+{% set text_input = text_input | default("How many singers do we have? | concert_singer | stadium : stadium_id, location, name, capacity, highest, lowest, average | singer : singer_id, name, country, song_name, song_release_year, age, is_male | concert : concert_id, concert_name, theme, stadium_id, year | singer_in_concert : concert_id, singer_id") %}
+{% set num_beams = num_beams | default(4) %}
+{% set max_length = max_length | default(512) %}
 
 repo_id = "{{ repo_id }}"

[EricLYunqi]

Fixed in v0.1.1, PyPI has been updated, so I will close this issue.