kanishka
/

smolm-autoreg-bpe-counterfactual-babylm-only_random_removal-seed_1024-1e-3

Text Generation

generated_from_trainer

Inference Endpoints

text-generation-inference

Model card Files Files and versions Metrics Training metrics Community

smolm-autoreg-bpe-counterfactual-babylm-only_random_removal-seed_1024-1e-3 / all_results.json

kanishka's picture

End of training

8ddd688 verified 3 months ago

raw history blame contribute delete

No virus

446 Bytes

	{
	"epoch": 20.0,
	"eval_accuracy": 0.41051635038682427,
	"eval_loss": 3.4108598232269287,
	"eval_runtime": 153.399,
	"eval_samples": 57919,
	"eval_samples_per_second": 377.571,
	"eval_steps_per_second": 5.9,
	"perplexity": 30.291278210076825,
	"train_loss": 3.0294145256159046,
	"train_runtime": 80976.6072,
	"train_samples": 594805,
	"train_samples_per_second": 146.908,
	"train_steps_per_second": 4.591
	}