SOC Estimation Using PyTorch and Pandas 98% accuracy on Dev set and entire dataset

Dataset used is confidential test data and cannot be shared

However, the repo att-ar/ecm_battery_simulation outputs data in the same structure as the dataset that was used:
time in seconds, current in Amperes, voltage in Volts, soc in percentage (regular SI units)

time	current	voltage	soc
0.0	0.0	3.65	95
1.0	-8.0	3.50	94.99
...	...	...	...

This repo contains a pytorch LSTM network that will be used to estimate State of Charge for a LFP Li-ion (LFP chemistry) cell from real time voltage, current and time data

The LSTM input will be voltage, current, and previous SOC points in a batch of windowed data of shape:
(G.batch_size, G.window_size, G.num_features)

The voltage, current and soc data will be from time: $$t - \text{windowsize} - 1 \rightarrow t - 1$$
The output should be the SOC prediction at time $t$ for each batch, the output shape should be (G.batch_size, 1)

98% accurate: