I am trying to export teknium/OpenHermes-2.5-Mistral-7B to ONNX,

This is my code:

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
import onnx
model_name = "teknium/OpenHermes-2.5-Mistral-7B"
model = AutoModelForCausalLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)
# Set the model to evaluation mode
model.eval()
# Prepare a dummy input for the export
dummy_input = tokenizer("This is a test input", return_tensors="pt").input_ids
# Export the model to ONNX
torch.onnx.export(
    model, 
    dummy_input, 
    "openhermes_onnx.onnx",
    input_names=["input_ids"], 
    output_names=["output"],
    dynamic_axes={"input_ids": {0: "batch_size", 1: "sequence_length"},
                  "output": {0: "batch_size", 1: "sequence_length"}},
    opset_version=14,
    do_constant_folding=True
)

It has generated the file (openhermes_onnx.onnx), but with 294 other Files with name similar to:

(model.embed_tokens.weight),(model.layers.0.post_attention_layernorm.weight),
(_model_layers.0_self_attn_rotary_emb_Constant_attr__value) and
(onnx__MatMul_10296)

Note: I have tried to load the model:

openhermes_onnx.onnx, but when I execute this code:

onnx_model_path = "openhermes_onnx.onnx"     
ort_session = ort.InferenceSession(onnx_model_path)

The kernel dies, and you have to restart the kernel.