def evaluate(args, model, tokenizer, device,data_type,prefix=""):
# if data_type == 'test':
# dataset,examples,features = load_and_cache_examples(args, tokenizer, data_type, output_examples=True,prefix = prefix)
# else:
dataset, examples, features = load_and_cache_examples(args, tokenizer, data_type = data_type, output_examples=True,prefix = prefix)
output_dir = os.path.join(args.output_dir,args.save_model_name)
if not os.path.exists(output_dir) and args.local_rank in [-1, 0]:
os.makedirs(output_dir)
# args.eval_batch_size = args.per_gpu_eval_batch_size * max(1, args.n_gpu)
# Note that DistributedSampler samples randomly
eval_sampler = SequentialSampler(dataset)
eval_dataloader = DataLoader(dataset, sampler=eval_sampler, batch_size=args.eval_batch_size)
# multi-gpu evaluate
if args.n_gpu > 1 and not isinstance(model, torch.nn.DataParallel):
model = torch.nn.DataParallel(model)
# Eval!
logger.info("***** Running evaluation {} *****".format(prefix))
logger.info(" Num examples = %d", len(dataset))
logger.info(" Batch size = %d", args.eval_batch_size)
all_results = []
start_time = timeit.default_timer()
for batch in tqdm(eval_dataloader, desc="Evaluating"):
model.eval()
batch = tuple(t.to(device) for t in batch)
with torch.no_grad():
inputs = {
"input_ids": batch[0],
"attention_mask": batch[1],
"token_type_ids": batch[2],
}
example_indices = batch[3]
outputs = model(**inputs)
# pdb.set_trace()
for i, example_index in enumerate(example_indices):
eval_feature = features[example_index.item()]
unique_id = int(eval_feature.unique_id)
output = [to_list(output[i]) for output in outputs]
# Some models (XLNet, XLM) use 5 arguments for their predictions, while the other "simpler"
# models only use two.
if len(output) >= 5:
start_logits = output[0]
start_top_index = output[1]
end_logits = output[2]
end_top_index = output[3]
cls_logits = output[4]
result = SquadResult(
unique_id,
start_logits,
end_logits,
start_top_index=start_top_index,
end_top_index=end_top_index,
cls_logits=cls_logits,
)
else:
start_logits, end_logits = output
result = SquadResult(unique_id, start_logits, end_logits)
all_results.append(result)
evalTime = timeit.default_timer() - start_time
logger.info(" Evaluation done in total %f secs (%f sec per example)", evalTime, evalTime / len(dataset))
# Compute predictions
output_prediction_file = os.path.join(output_dir, "predictions_{}.json".format(prefix))
output_nbest_file = os.path.join(output_dir, "nbest_predictions_{}.json".format(prefix))
output_null_log_odds_file = None
predictions = compute_predictions_logits(
examples,
features,
all_results,
args.n_best_size,
args.max_answer_length,
args.do_lower_case,
output_prediction_file,
output_nbest_file,
output_null_log_odds_file,
args.verbose_logging,
False,
args.null_score_diff_threshold,
tokenizer,
)
# Compute the F1 and exact scores.
if data_type == 'test':
return
dev_dir = os.path.join(args.data_dir,args.dev_file)
dev = json.load(open(dev_dir,'r'))
prediction = json.load(open(output_prediction_file))
F1, EM, TOTAL, SKIP = Eval.evaluate(dev,prediction)
return F1, EM
def evaluate(args, model, tokenizer, prefix=""):
dataset, examples, features = load_and_cache_examples(args,
tokenizer,
evaluate=True,
output_examples=True)
if not os.path.exists(args.output_dir) and args.local_rank in [-1, 0]:
os.makedirs(args.output_dir)
args.eval_batch_size = args.per_gpu_eval_batch_size * max(1, args.n_gpu)
# Note that DistributedSampler samples randomly
eval_sampler = SequentialSampler(dataset)
eval_dataloader = DataLoader(dataset,
sampler=eval_sampler,
batch_size=args.eval_batch_size)
# multi-gpu evaluate
if args.n_gpu > 1 and not isinstance(model, torch.nn.DataParallel):
model = torch.nn.DataParallel(model)
# Eval!
logger.info("***** Running evaluation {} *****".format(prefix))
logger.info(" Num examples = %d", len(dataset))
logger.info(" Batch size = %d", args.eval_batch_size)
all_results = []
start_time = timeit.default_timer()
for batch in tqdm(eval_dataloader, desc="Evaluating"):
model.eval()
batch = tuple(t.to(args.device) for t in batch)
with torch.no_grad():
inputs = {
"input_ids": batch[0],
"attention_mask": batch[1],
"token_type_ids": batch[2],
}
if args.use_context_keyword:
inputs["keyword_type_ids"] = batch[6]
else:
inputs["keyword_type_ids"] = batch[7]
if args.model_type in [
"xlm", "roberta", "distilbert", "camembert"
]:
del inputs["token_type_ids"]
example_indices = batch[3]
# XLNet and XLM use more arguments for their predictions
if args.model_type in ["xlnet", "xlm"]:
inputs.update({"cls_index": batch[4], "p_mask": batch[5]})
# for lang_id-sensitive xlm models
if hasattr(model, "config") and hasattr(
model.config, "lang2id"):
inputs.update({
"langs":
(torch.ones(batch[0].shape, dtype=torch.int64) *
args.lang_id).to(args.device)
})
outputs = model(**inputs)
for i, example_index in enumerate(example_indices):
eval_feature = features[example_index.item()]
unique_id = int(eval_feature.unique_id)
output = [to_list(output[i]) for output in outputs]
# Some models (XLNet, XLM) use 5 arguments for their predictions, while the other "simpler"
# models only use two.
if len(output) >= 5:
start_logits = output[0]
start_top_index = output[1]
end_logits = output[2]
end_top_index = output[3]
cls_logits = output[4]
result = SquadResult(
unique_id,
start_logits,
end_logits,
start_top_index=start_top_index,
end_top_index=end_top_index,
cls_logits=cls_logits,
)
else:
start_logits, end_logits = output
result = SquadResult(unique_id, start_logits, end_logits)
all_results.append(result)
evalTime = timeit.default_timer() - start_time
logger.info(" Evaluation done in total %f secs (%f sec per example)",
evalTime, evalTime / len(dataset))
# Compute predictions
output_prediction_file = os.path.join(args.output_result_dir,
"predictions_{}.json".format(prefix))
output_nbest_file = os.path.join(
args.output_result_dir, "nbest_predictions_{}.json".format(prefix))
output_all_logit_file = os.path.join(
args.output_result_dir, "logit_predictions_{}.json".format(prefix))
if args.version_2_with_negative:
output_null_log_odds_file = os.path.join(
args.output_result_dir, "null_odds_{}.json".format(prefix))
else:
output_null_log_odds_file = None
# XLNet and XLM use a more complex post-processing procedure
if args.model_type in ["xlnet", "xlm"]:
start_n_top = model.config.start_n_top if hasattr(
model, "config") else model.module.config.start_n_top
end_n_top = model.config.end_n_top if hasattr(
model, "config") else model.module.config.end_n_top
predictions = compute_predictions_log_probs(
examples,
features,
all_results,
args.n_best_size,
args.max_answer_length,
output_prediction_file,
output_nbest_file,
output_null_log_odds_file,
start_n_top,
end_n_top,
args.version_2_with_negative,
tokenizer,
args.verbose_logging,
)
else:
if args.output_all_logit:
predictions = compute_predictions_logits_and_output_all_logits(
examples,
features,
all_results,
args.n_best_size,
args.max_answer_length,
args.do_lower_case,
output_prediction_file,
output_nbest_file,
output_all_logit_file,
output_null_log_odds_file,
args.verbose_logging,
args.version_2_with_negative,
args.null_score_diff_threshold,
tokenizer,
)
else:
predictions = compute_predictions_logits(
examples,
features,
all_results,
args.n_best_size,
args.max_answer_length,
args.do_lower_case,
output_prediction_file,
output_nbest_file,
output_null_log_odds_file,
args.verbose_logging,
args.version_2_with_negative,
args.null_score_diff_threshold,
tokenizer,
)
# Compute the F1 and exact scores.
results = squad_evaluate(examples, predictions)
return results
def evaluateQA(model, corpus, task, path):
dataset, examples, features = corpus.dataset, corpus.examples, corpus.features
tokenizer = corpus.tokenizer
# Note that DistributedSampler samples randomly
eval_sampler = SequentialSampler(dataset)
eval_dataloader = DataLoader(dataset, sampler=eval_sampler, batch_size=12)
all_results = []
for batch in eval_dataloader:
model.eval()
with torch.no_grad():
inputs = {
"input_ids": batch[0],
"attention_mask": batch[1],
"token_type_ids": batch[2],
"answer_start": None,
"answer_end": None,
}
example_indices = batch[3]
outputs = model('qa', inputs)
for i, example_index in enumerate(example_indices):
eval_feature = features[example_index.item()]
unique_id = int(eval_feature.unique_id)
output = [to_list(output[i]) for output in outputs]
start_logits, end_logits = output
result = SquadResult(unique_id, start_logits, end_logits)
all_results.append(result)
# Compute predictions
output_prediction_file = os.path.join(path,
"predictions_{}.json".format(task))
output_nbest_file = os.path.join(path,
"nbest_predictions_{}.json".format(task))
output_null_log_odds_file = None
predictions = compute_predictions_logits(
examples,
features,
all_results,
20,
30,
True,
output_prediction_file,
output_nbest_file,
output_null_log_odds_file,
False,
False,
0.0,
tokenizer,
)
# Compute the F1 and exact scores.
results = squad_evaluate(examples, predictions)
return results