Model does not export properly

[mdaftary]

(PS: I am using google colab to do all of this)
So, I was trying to use the sparseml zoo and train one of the models (https://sparsezoo.neuralmagic.com/models/yolov8-n-coco-pruned49_quantized?hardware=deepsparse-c6i.12xlarge) with my custom dataset. I trained it and exported it but my finetuned model.onnx file size is ~4.7 MB, while the original pretrained model.onnx file size is ~3.5 MB. I thought the size should remain the same after finetuning?
Also I tried benchmarking it on the CPU using deepsparse.benchmark and I get these results:

2024-06-26 03:39:24 deepsparse.benchmark.helpers INFO Thread pinning to cores enabled
DeepSparse, Copyright 2021-present / Neuralmagic, Inc. version: 1.7.1 COMMUNITY | (3904e8ec) (release) (optimized) (system=avx2, binary=avx2)
2024-06-26 03:39:25 deepsparse.benchmark.benchmark_model INFO deepsparse.engine.Engine:
onnx_file_path: /content/drive/MyDrive/sparse_model_runs/detect/sparse49_56epochs/deployment/model.onnx
batch_size: 1
num_cores: 1
num_streams: 1
scheduler: Scheduler.default
fraction_of_supported_ops: 0.0
cpu_avx_type: avx2
cpu_vnni: False
2024-06-26 03:39:25 deepsparse.utils.onnx INFO Generating input 'images', type = uint8, shape = [1, 3, 640, 640]
2024-06-26 03:39:25 deepsparse.benchmark.benchmark_model INFO Starting 'singlestream' performance measurements for 10 seconds
Original Model Path: /content/drive/MyDrive/sparse_model_runs/detect/sparse49_56epochs/deployment/model.onnx
Batch Size: 1
Scenario: sync
Throughput (items/sec): 3.6589
Latency Mean (ms/batch): 273.2898
Latency Median (ms/batch): 235.4801
Latency Std (ms/batch): 58.9786
Iterations: 37

Why is the fraction_of_supported_ops = 0.0

When exporting the model, I get these warnings:
2024-06-26 01:49:44.989820: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2024-06-26 01:49:47.102603: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT

Could it be related to this? because I also tried exporting the model.pt file of the original pretrained model and the exported model.onnx file was still around ~4.7 MB, even though in the documentation it is ~3.5 MB.

I tried benchmarking my exported version of the original pretrained model and comparing it to the benchmark of the pretrained model.onnx file and again I get similar differences. The fraction_of_supported_ops = 0.0 for my exported version but it is =1 for the model.onnx file from the website.

Would appreciate any help and can provide any additional information.