--- license: other library_name: peft tags: - trl - sft - generated_from_trainer base_model: google/gemma-2b model-index: - name: Gemma-2b-chat results: [] --- # Gemma-2b-chat This model is a fine-tuned version of [google/gemma-2b](https://huggingface.co./google/gemma-2b) on the None dataset. It achieves the following results on the evaluation set: - Loss: 2.3272 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 1 - eval_batch_size: 1 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - num_epochs: 3 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:-----:|:---------------:| | 2.974 | 0.02 | 100 | 2.9419 | | 3.1998 | 0.04 | 200 | 2.9117 | | 2.9839 | 0.06 | 300 | 2.8745 | | 2.9379 | 0.08 | 400 | 2.8328 | | 2.9392 | 0.1 | 500 | 2.7909 | | 2.7735 | 0.12 | 600 | 2.7461 | | 2.7835 | 0.14 | 700 | 2.7075 | | 2.7794 | 0.16 | 800 | 2.6695 | | 2.7355 | 0.18 | 900 | 2.6363 | | 2.7805 | 0.2 | 1000 | 2.6104 | | 2.556 | 0.22 | 1100 | 2.5871 | | 2.596 | 0.24 | 1200 | 2.5652 | | 2.5838 | 0.26 | 1300 | 2.5432 | | 2.6598 | 0.28 | 1400 | 2.5247 | | 2.5475 | 0.3 | 1500 | 2.5100 | | 2.4459 | 0.31 | 1600 | 2.4954 | | 2.502 | 0.33 | 1700 | 2.4829 | | 2.557 | 0.35 | 1800 | 2.4702 | | 2.4944 | 0.37 | 1900 | 2.4604 | | 2.4774 | 0.39 | 2000 | 2.4528 | | 2.4287 | 0.41 | 2100 | 2.4453 | | 2.5386 | 0.43 | 2200 | 2.4381 | | 2.363 | 0.45 | 2300 | 2.4322 | | 2.514 | 0.47 | 2400 | 2.4272 | | 2.413 | 0.49 | 2500 | 2.4225 | | 2.4667 | 0.51 | 2600 | 2.4176 | | 2.4724 | 0.53 | 2700 | 2.4128 | | 2.3949 | 0.55 | 2800 | 2.4084 | | 2.4822 | 0.57 | 2900 | 2.4044 | | 2.4556 | 0.59 | 3000 | 2.4009 | | 2.4067 | 0.61 | 3100 | 2.3977 | | 2.3911 | 0.63 | 3200 | 2.3947 | | 2.3446 | 0.65 | 3300 | 2.3923 | | 2.3358 | 0.67 | 3400 | 2.3891 | | 2.3213 | 0.69 | 3500 | 2.3867 | | 2.4041 | 0.71 | 3600 | 2.3840 | | 2.4759 | 0.73 | 3700 | 2.3818 | | 2.4622 | 0.75 | 3800 | 2.3801 | | 2.3512 | 0.77 | 3900 | 2.3778 | | 2.3653 | 0.79 | 4000 | 2.3760 | | 2.3455 | 0.81 | 4100 | 2.3744 | | 2.4364 | 0.83 | 4200 | 2.3724 | | 2.2805 | 0.85 | 4300 | 2.3706 | | 2.5448 | 0.87 | 4400 | 2.3681 | | 2.3061 | 0.89 | 4500 | 2.3674 | | 2.2572 | 0.9 | 4600 | 2.3657 | | 2.3259 | 0.92 | 4700 | 2.3645 | | 2.4078 | 0.94 | 4800 | 2.3633 | | 2.3841 | 0.96 | 4900 | 2.3618 | | 2.5439 | 0.98 | 5000 | 2.3604 | | 2.4556 | 1.0 | 5100 | 2.3593 | | 2.3752 | 1.02 | 5200 | 2.3582 | | 2.3415 | 1.04 | 5300 | 2.3567 | | 2.2824 | 1.06 | 5400 | 2.3555 | | 2.3748 | 1.08 | 5500 | 2.3541 | | 2.2535 | 1.1 | 5600 | 2.3534 | | 2.3277 | 1.12 | 5700 | 2.3530 | | 2.394 | 1.14 | 5800 | 2.3518 | | 2.4876 | 1.16 | 5900 | 2.3511 | | 2.4705 | 1.18 | 6000 | 2.3503 | | 2.4394 | 1.2 | 6100 | 2.3499 | | 2.3898 | 1.22 | 6200 | 2.3488 | | 2.3789 | 1.24 | 6300 | 2.3483 | | 2.4315 | 1.26 | 6400 | 2.3472 | | 2.4065 | 1.28 | 6500 | 2.3463 | | 2.3331 | 1.3 | 6600 | 2.3456 | | 2.3415 | 1.32 | 6700 | 2.3452 | | 2.3433 | 1.34 | 6800 | 2.3448 | | 2.337 | 1.36 | 6900 | 2.3434 | | 2.4492 | 1.38 | 7000 | 2.3425 | | 2.3757 | 1.4 | 7100 | 2.3419 | | 2.4124 | 1.42 | 7200 | 2.3412 | | 2.2778 | 1.44 | 7300 | 2.3408 | | 2.3127 | 1.46 | 7400 | 2.3401 | | 2.2558 | 1.48 | 7500 | 2.3398 | | 2.4419 | 1.49 | 7600 | 2.3394 | | 2.3052 | 1.51 | 7700 | 2.3388 | | 2.3212 | 1.53 | 7800 | 2.3387 | | 2.3989 | 1.55 | 7900 | 2.3376 | | 2.3201 | 1.57 | 8000 | 2.3372 | | 2.4111 | 1.59 | 8100 | 2.3364 | | 2.3243 | 1.61 | 8200 | 2.3361 | | 2.3158 | 1.63 | 8300 | 2.3360 | | 2.3065 | 1.65 | 8400 | 2.3357 | | 2.3627 | 1.67 | 8500 | 2.3353 | | 2.4604 | 1.69 | 8600 | 2.3348 | | 2.2451 | 1.71 | 8700 | 2.3346 | | 2.3559 | 1.73 | 8800 | 2.3342 | | 2.4832 | 1.75 | 8900 | 2.3338 | | 2.5064 | 1.77 | 9000 | 2.3335 | | 2.2961 | 1.79 | 9100 | 2.3336 | | 2.4394 | 1.81 | 9200 | 2.3334 | | 2.4337 | 1.83 | 9300 | 2.3332 | | 2.2984 | 1.85 | 9400 | 2.3328 | | 2.2544 | 1.87 | 9500 | 2.3325 | | 2.4421 | 1.89 | 9600 | 2.3321 | | 2.2737 | 1.91 | 9700 | 2.3322 | | 2.4483 | 1.93 | 9800 | 2.3319 | | 2.4371 | 1.95 | 9900 | 2.3314 | | 2.3184 | 1.97 | 10000 | 2.3312 | | 2.2936 | 1.99 | 10100 | 2.3308 | | 2.432 | 2.01 | 10200 | 2.3304 | | 2.3306 | 2.03 | 10300 | 2.3301 | | 2.3926 | 2.05 | 10400 | 2.3301 | | 2.358 | 2.07 | 10500 | 2.3300 | | 2.341 | 2.08 | 10600 | 2.3298 | | 2.3886 | 2.1 | 10700 | 2.3297 | | 2.2559 | 2.12 | 10800 | 2.3296 | | 2.4121 | 2.14 | 10900 | 2.3294 | | 2.3301 | 2.16 | 11000 | 2.3292 | | 2.2807 | 2.18 | 11100 | 2.3290 | | 2.3028 | 2.2 | 11200 | 2.3288 | | 2.2957 | 2.22 | 11300 | 2.3289 | | 2.296 | 2.24 | 11400 | 2.3289 | | 2.248 | 2.26 | 11500 | 2.3288 | | 2.3639 | 2.28 | 11600 | 2.3286 | | 2.4383 | 2.3 | 11700 | 2.3284 | | 2.2921 | 2.32 | 11800 | 2.3282 | | 2.4594 | 2.34 | 11900 | 2.3282 | | 2.4243 | 2.36 | 12000 | 2.3280 | | 2.344 | 2.38 | 12100 | 2.3280 | | 2.3063 | 2.4 | 12200 | 2.3279 | | 2.3875 | 2.42 | 12300 | 2.3280 | | 2.3502 | 2.44 | 12400 | 2.3278 | | 2.3034 | 2.46 | 12500 | 2.3278 | | 2.4234 | 2.48 | 12600 | 2.3277 | | 2.2829 | 2.5 | 12700 | 2.3277 | | 2.3965 | 2.52 | 12800 | 2.3277 | | 2.4046 | 2.54 | 12900 | 2.3274 | | 2.3374 | 2.56 | 13000 | 2.3274 | | 2.1988 | 2.58 | 13100 | 2.3274 | | 2.3893 | 2.6 | 13200 | 2.3274 | | 2.3621 | 2.62 | 13300 | 2.3273 | | 2.2888 | 2.64 | 13400 | 2.3273 | | 2.3928 | 2.66 | 13500 | 2.3273 | | 2.3523 | 2.68 | 13600 | 2.3272 | | 2.3158 | 2.69 | 13700 | 2.3273 | | 2.3453 | 2.71 | 13800 | 2.3273 | | 2.3113 | 2.73 | 13900 | 2.3272 | | 2.3878 | 2.75 | 14000 | 2.3272 | | 2.3361 | 2.77 | 14100 | 2.3273 | | 2.2343 | 2.79 | 14200 | 2.3273 | | 2.2963 | 2.81 | 14300 | 2.3271 | | 2.252 | 2.83 | 14400 | 2.3272 | | 2.4307 | 2.85 | 14500 | 2.3272 | | 2.2778 | 2.87 | 14600 | 2.3272 | | 2.3832 | 2.89 | 14700 | 2.3272 | | 2.3611 | 2.91 | 14800 | 2.3272 | | 2.3556 | 2.93 | 14900 | 2.3271 | | 2.3712 | 2.95 | 15000 | 2.3272 | | 2.3667 | 2.97 | 15100 | 2.3272 | | 2.3816 | 2.99 | 15200 | 2.3272 | ### Framework versions - PEFT 0.9.0 - Transformers 4.38.2 - Pytorch 2.1.2 - Datasets 2.17.1 - Tokenizers 0.15.2