Error on privateGPT.py execution

[shameforest]

I managed to use ingest.py and load the txt files.
Afterwards I try to execture privateGPT.py and get the following error message (every time):

Using embedded DuckDB with persistence: data will be stored in: db
Traceback (most recent call last):
File "c:\privateGPT-main\privateGPT.py", line 75, in
main()
File "c:\privateGPT-main\privateGPT.py", line 35, in main
llm = GPT4All(model=model_path, n_ctx=model_n_ctx, backend='gptj', callbacks=callbacks, verbose=False)
File "pydantic\main.py", line 341, in pydantic.main.BaseModel.init
pydantic.error_wrappers.ValidationError: 1 validation error for GPT4All
root
Failed to retrieve model (type=value_error)

---

• Windows 11
• Python version 3.10.9

---

my env file looks like this:

MODEL_TYPE=GPT4All
PERSIST_DIRECTORY=db
MODEL_PATH=models\ggml-gpt4all-j-v1.3-groovy.bin
MODEL_N_CTX=1000
EMBEDDINGS_MODEL_NAME=distiluse-base-multilingual-cased-v2

[sandyrs9421]

i too see the same error . any luck on how we can resolve this ?

my env file -
PERSIST_DIRECTORY=db
MODEL_TYPE=GPT4All
MODEL_PATH=models/ggml-gpt4all-j-v1.3-groovy.bin
EMBEDDINGS_MODEL_NAME=/Users/FBT/Desktop/Projects/privategpt/privateGPT/models/ggml-model-q4_0.bin
MODEL_N_CTX=1000

[christopherpickering]

Did you download the model in this path models\ggml-gpt4all-j-v1.3-groovy.bin?

[shameforest]

Yes it is in that path, I copied the relative path to the file and inserted it in the env file, like described above.

[UserB-tm]

I'm also able to ingest but can't seem's to stick at model load.

[christopherpickering]

Does the groovy model work? I removed the "backend=gptj" in the privateGPT.py file to get the model to load.

[shameforest]

I removed backend=gptj now. This is my error message:

Traceback (most recent call last):
File "c:\privateGPT-main\privateGPT.py", line 75, in
main()
File "c:\privateGPT-main\privateGPT.py", line 35, in main
llm = GPT4All(model=model_path, n_ctx=model_n_ctx, callbacks=callbacks, verbose=False)
File "pydantic\main.py", line 341, in pydantic.main.BaseModel.init
pydantic.error_wrappers.ValidationError: 1 validation error for GPT4All
root
No corresponding model for provided filename models\ggml-gpt4all-j-v1.3-groovy.bin.
If this is a custom model, make sure to specify a valid model_type.
(type=value_error)

---

in my models folder are these two files:
models\ggml-gpt4all-j-v1.3-groovy.bin
models\ggml-model-q4_0.bin

[Andrew-Mayorga]

make sure the llm is inside the models folder

[shameforest]

You are speaking of:
models\ggml-gpt4all-j-v1.3-groovy.bin
models\ggml-model-q4_0.bin

right?

They are both in the models folder, in the real file system (C:\privateGPT-main\models) and inside Visual Studio Code (models\ggml-gpt4all-j-v1.3-groovy.bin) aswell.
Is there anything else that could be the problem?

[HenryUn]

I guess the error might cause by "model" which it should be "model_path"

The error show:
File "c:\privateGPT-main\privateGPT.py", line 35, in main
llm = GPT4All(model=model_path, n_ctx=model_n_ctx, callbacks=callbacks, verbose=False)

I checked the privateGPT.py file. it only has defined :model_path = os.environ.get('MODEL_PATH')
sso, I guess that the code should be

llm = GPT4All(model_path=model_path, n_ctx=model_n_ctx, callbacks=callbacks, verbose=False)

Please do feel free to correct any of my mistakes.. thanks.

[shameforest]

I replaced your code with my line 35, but it just gets this error:

Using embedded DuckDB with persistence: data will be stored in: db
Traceback (most recent call last):
File "c:\privateGPT-main\privateGPT.py", line 75, in
main()
File "c:\privateGPT-main\privateGPT.py", line 35, in main
llm = GPT4All(model_path=model_path, n_ctx=model_n_ctx, callbacks=callbacks, verbose=False)
File "pydantic\main.py", line 339, in pydantic.main.BaseModel.init
File "pydantic\main.py", line 1102, in pydantic.main.validate_model
File "C:\Users\Amin\AppData\Local\Programs\Python\Python310\lib\site-packages\langchain\llms\gpt4all.py", line 135, in validate_environment
full_path = values["model"]
KeyError: 'model'

[buhajbej]

Same here. The "model" to "model_path" helped with the privateGPT file error, but now stuck with the gpt4all file error.

[bdbais]

When I try to run privateGPT: I have last version of python... how to run on compatible version mode?

python3 privateGPT.py
File "/home/privateGPT/privateGPT/privateGPT.py", line 32
match model_type:
^
SyntaxError: invalid syntax

python3 --version
Python 3.9.7

uname -an
Linux hostname 5.13.0-27-generic #29-Ubuntu SMP Wed Jan 12 17:36:47 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux

[Fridry]

I had the same error with this .env file:

PERSIST_DIRECTORY=db
MODEL_TYPE=GPT4All
MODEL_PATH=models/ggml-gpt4all-j-v1.3-groovy.bin
EMBEDDINGS_MODEL_NAME=/Users/FBT/Desktop/Projects/privategpt/privateGPT/models/ggml-model-q4_0.bin
MODEL_N_CTX=1000

I just replace the models on MODEL_TYPE for source_documents. My MODEL_TYPE is like this now:
MODEL_TYPE: source_documents/ggml-gpt4all-j-v1.3-groovy.bin

[geller6980]

I had the same error with this .env file:

PERSIST_DIRECTORY=db MODEL_TYPE=GPT4All MODEL_PATH=models/ggml-gpt4all-j-v1.3-groovy.bin EMBEDDINGS_MODEL_NAME=/Users/FBT/Desktop/Projects/privategpt/privateGPT/models/ggml-model-q4_0.bin MODEL_N_CTX=1000

I just replace the models on MODEL_TYPE for source_documents. My MODEL_TYPE is like this now: MODEL_TYPE: source_documents/ggml-gpt4all-j-v1.3-groovy.bin

This worked perfectly!

I also commented out the lines under "Print the relevant sources used for the answer" to shorten the text.

Is there a way to speed up the response time?

[hansvanzutphen]

Found it. Don't use a \ in the path. So replace
MODEL_PATH=models\ggml-gpt4all-j-v1.3-groovy.bin
by
MODEL_PATH=models/ggml-gpt4all-j-v1.3-groovy.bin

[mvange]

Solved:

#89

[m3hrdadpourkheiri]

Hi dear
I run ingesty.py by python 3.9.13 and work correctly
but when run privateGPT.py return this error
C:\Users\Mehrdad\Desktop\AI\newpgpt\privateGPT>python privateGPT.py
File "C:\Users\Mehrdad\Desktop\AI\newpgpt\privateGPT\privateGPT.py", line 34
match model_type:
^
SyntaxError: invalid syntax

please help me

[naveeagrawal]

@m3hrdadpourkheiri match statement would work only in Python 3.10 above.