learning_rate: 5e-05
train_batch_size: 16
eval_batch_size: 16
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: constant
lr_scheduler_warmup_steps: 0.2
num_epochs: 3

Training Loss	Epoch	Step	Validation Loss	Per Phonemes	Per Hams Or Jahr	Per Shidda Or Rakhawa	Per Tafkheem Or Taqeeq	Per Itbaq	Per Safeer	Per Qalqla	Per Tikraar	Per Tafashie	Per Istitala	Per Ghonna	Average Per
2.1346	0.6000	7279	0.4810	0.1782	0.1304	0.1829	0.4285	0.1592	0.1340	0.3205	0.1452	0.1766	0.1953	0.1024	0.1958
0.3336	1.2001	14558	0.2327	0.0901	0.0481	0.0756	0.1104	0.0271	0.0482	0.0376	0.0285	0.0249	0.0216	0.0406	0.0503
0.1949	1.8001	21837	0.1582	0.0597	0.0297	0.0493	0.0712	0.0181	0.0375	0.0260	0.0189	0.0187	0.0145	0.0281	0.0338
0.1409	2.4001	29116	0.1178	0.0537	0.0282	0.0436	0.0671	0.0257	0.0355	0.0266	0.0213	0.0203	0.0240	0.0290	0.0341
0.1409	3.0	36393	0.0994	0.0476	0.0239	0.0356	0.0609	0.0225	0.0295	0.0245	0.0196	0.0162	0.0224	0.0263	0.0299

AbdullahHedeya/h4_f1024_l16_d256_no_kd_3_epochs

h4_f1024_l16_d256_no_kd_3_epochs