Meta-Llama-3-8B-Generator-logging

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2138

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
1.456	0.3287	20	0.7801
0.662	0.6574	40	0.5845
0.5359	0.9861	60	0.4810
0.3957	1.3148	80	0.2857
0.272	1.6436	100	0.2681
0.2566	1.9723	120	0.2467
0.2427	2.3010	140	0.2387
0.2371	2.6297	160	0.2361
0.2337	2.9584	180	0.2312
0.2302	3.2871	200	0.2289
0.228	3.6158	220	0.2270
0.2268	3.9445	240	0.2252
0.2239	4.2732	260	0.2232
0.2223	4.6020	280	0.2232
0.2215	4.9307	300	0.2215
0.2194	5.2594	320	0.2193
0.219	5.5881	340	0.2200
0.2186	5.9168	360	0.2182
0.2165	6.2455	380	0.2177
0.2164	6.5742	400	0.2172
0.216	6.9029	420	0.2168
0.215	7.2316	440	0.2162
0.2143	7.5603	460	0.2160
0.2136	7.8891	480	0.2150
0.213	8.2178	500	0.2148
0.2127	8.5465	520	0.2145
0.2124	8.8752	540	0.2141
0.2113	9.2039	560	0.2139
0.2113	9.5326	580	0.2138
0.2115	9.8613	600	0.2138

Framework versions

• PEFT 0.10.0
• Transformers 4.42.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.14.7
• Tokenizers 0.19.1