Meta-Llama-3-8B_magiccoder_default

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

• Loss: 1.2697

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: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.02
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.2592	0.0259	4	1.4263
1.4281	0.0518	8	1.4063
1.3795	0.0777	12	1.3824
1.3751	0.1036	16	1.3937
1.4053	0.1296	20	1.3523
1.2927	0.1555	24	1.3474
1.3619	0.1814	28	1.3529
1.3533	0.2073	32	1.3629
1.3627	0.2332	36	1.3636
1.4408	0.2591	40	1.3531
1.3744	0.2850	44	1.3395
1.2658	0.3109	48	1.3364
1.3364	0.3368	52	1.3400
1.3765	0.3628	56	1.3391
1.3427	0.3887	60	1.3370
1.3975	0.4146	64	1.3329
1.2595	0.4405	68	1.3325
1.3291	0.4664	72	1.3312
1.2702	0.4923	76	1.3323
1.3527	0.5182	80	1.3213
1.2799	0.5441	84	1.3154
1.3082	0.5700	88	1.3099
1.4042	0.5960	92	1.3089
1.2221	0.6219	96	1.3048
1.3079	0.6478	100	1.3017
1.2165	0.6737	104	1.2970
1.239	0.6996	108	1.2941
1.2528	0.7255	112	1.2877
1.2932	0.7514	116	1.2859
1.2762	0.7773	120	1.2804
1.2914	0.8032	124	1.2791
1.2835	0.8291	128	1.2755
1.2735	0.8551	132	1.2731
1.2264	0.8810	136	1.2722
1.2637	0.9069	140	1.2713
1.2133	0.9328	144	1.2704
1.2379	0.9587	148	1.2699
1.2131	0.9846	152	1.2697

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1