Python Machine learning
Budget: $250 – $750 USD
I am looking for a Python machine learning expert to assist me with my project. Here are the details:
Specific Dataset:
- You will get a 2 sample files that are 1GB in size
Preferred Machine Learning Model:
- You have to convert the current python script used to the new usecase. The new usecase and documentation will be delivered on accepting the challenge.
- Attached you will find a sample of the custom model that is used for mnist. We will NOT use mnist, this is a example so you know which script you would have to edit for the new task. Everything has to be changed to Tinygrad Tensor calculations instead of numpy. In this way we can run all action on the GPU.
Desired Outcome:
- You provide 2 scripts (all abreviated from my script) but all changed to tinygrad python library.
-- Script 1: training.py
--- You create the specified custom dataloader
--- You change the script so that it can 100% use the GPU's parallelism via the tinygrad ml library
--- Batchsize and auto savepoints of the model after steps X
--Script 2: inference.py
--- variable to chose which model to load
--- load the model and do the inference and output the result to the console with a normal print statement.
Ideal Skills and Experience:
- Proficiency in Python and machine learning algorithms
- Experience with the Tinygrad library
- Knowledge of custom dataloader and inference techniques
- Ability to work with specific datasets and deliver accurate predictions
- Strong problem-solving and analytical skills
If you have the necessary skills and experience, please submit your proposal.
Specific Dataset:
- You will get a 2 sample files that are 1GB in size
Preferred Machine Learning Model:
- You have to convert the current python script used to the new usecase. The new usecase and documentation will be delivered on accepting the challenge.
- Attached you will find a sample of the custom model that is used for mnist. We will NOT use mnist, this is a example so you know which script you would have to edit for the new task. Everything has to be changed to Tinygrad Tensor calculations instead of numpy. In this way we can run all action on the GPU.
Desired Outcome:
- You provide 2 scripts (all abreviated from my script) but all changed to tinygrad python library.
-- Script 1: training.py
--- You create the specified custom dataloader
--- You change the script so that it can 100% use the GPU's parallelism via the tinygrad ml library
--- Batchsize and auto savepoints of the model after steps X
--Script 2: inference.py
--- variable to chose which model to load
--- load the model and do the inference and output the result to the console with a normal print statement.
Ideal Skills and Experience:
- Proficiency in Python and machine learning algorithms
- Experience with the Tinygrad library
- Knowledge of custom dataloader and inference techniques
- Ability to work with specific datasets and deliver accurate predictions
- Strong problem-solving and analytical skills
If you have the necessary skills and experience, please submit your proposal.