💡 All-in-one AI framework for semantic search, LLM orchestration and language model workflows
Python agentsaiai-agentsembeddingsinformation-retrievallanguage-modellarge-language-modelsllmnlppythonragretrieval-augmented-generationsearchsearch-enginesemantic-searchsentence-embeddingstransformerstxtaivector-databasevector-search
Pass evaluation metrics to underlying Trainer, closes #398
davidmezzettia8c8bd88ae5f
diff --git a/src/python/txtai/pipeline/train/hftrainer.py b/src/python/txtai/pipeline/train/hftrainer.pyindex 48fd43f1..1c89e0f3 100644--- a/src/python/txtai/pipeline/train/hftrainer.py+++ b/src/python/txtai/pipeline/train/hftrainer.py@@ -19,7 +19,9 @@ class HFTrainer(Tensors):Trains a new Hugging Face Transformer model using the Trainer framework."""- def __call__(self, base, train, validation=None, columns=None, maxlength=None, stride=128, task="text-classification", prefix=None, **args):+ def __call__(+ self, base, train, validation=None, columns=None, maxlength=None, stride=128, task="text-classification", prefix=None, metrics=None, **args+ ):"""Builds a new model using arguments.@@ -32,6 +34,7 @@ class HFTrainer(Tensors):stride: chunk size for splitting data for QA taskstask: optional model task or category, determines the model type, defaults to "text-classification"prefix: optional source prefix+ metrics: optional function that computes and returns a dict of evaluation metricsargs: training argumentsReturns:@@ -72,7 +75,13 @@ class HFTrainer(Tensors):# Build trainertrainer = Trainer(- model=model, tokenizer=tokenizer, data_collator=collator, args=args, train_dataset=train, eval_dataset=validation if validation else None+ model=model,+ tokenizer=tokenizer,+ data_collator=collator,+ args=args,+ train_dataset=train,+ eval_dataset=validation if validation else None,+ compute_metrics=metrics,)# Run training