Specify data type in CountVectorzier

**What happened**:

I have the data in the following format:
```
0 Satellite TV|Golf Course|Airport Shuttle|Cosme...
1 Satellite TV|Cosmetic Mirror|Safe (Hotel)|Tele...
2 Satellite TV|Cosmetic Mirror|Safe (Hotel)|Tele...
3 Satellite TV|Sailing|Cosmetic Mirror|Telephone...
4 Satellite TV|Sailing|Diving|Cosmetic Mirror|Sa...
```

```
vect = CountVectorizer(tokenizer=lambda x: x.split("|"))
tf_df = vect.fit_transform(item_data['properties'])
```

**What you expected to happen**:

I want to use `CountVectorizer` to get the dataframe with the corresponding columns

Seems I need to specify the output type somehow. But I don't see CountVectorizer having an interface to specify metadata

```
ValueError: Metadata inference failed in `_count_vectorizer_transform`.

You have supplied a custom function and Dask is unable to
determine the type of output that that function returns.

To resolve this please provide a meta= keyword.
The docstring of the Dask function you ran should have more information.
```

**Minimal Complete Verifiable Example**:

```python
# Put your MCVE code here
```

**Anything else we need to know?**:

**Environment**:

- Dask version:
```
dask 2021.1.1 pyhd3eb1b0_0
dask-core 2021.1.1 pyhd3eb1b0_0
dask-glm 0.2.0 py38_0
dask-ml 1.8.0 pyhd3eb1b0_0
```
- Python version: Python 3.8.13
- Operating System: Ubuntu 20.04
- Install method (conda, pip, source): conda

Cluster Dump State: