Sponsored by APIMart
Toolify Toolify logo
AI Models
albert-xlarge-v2-squad-v2

ktrapeznikov / albert-xlarge-v2-squad-v2

Run on huggingface.co
huggingface.co
Total runs: 109
24-hour runs: 0
7-day runs: 30
30-day runs: -39
Model's Last Updated: December 12 2020
question-answering

Introduction of albert-xlarge-v2-squad-v2

Model Details of albert-xlarge-v2-squad-v2

Model

albert-xlarge-v2 fine-tuned on SQuAD V2 using run_squad.py

Training Parameters

Trained on 4 NVIDIA GeForce RTX 2080 Ti 11Gb 

BASE_MODEL=albert-xlarge-v2
python run_squad.py \
  --version_2_with_negative \
  --model_type albert \
  --model_name_or_path $BASE_MODEL \
  --output_dir $OUTPUT_MODEL \
  --do_eval \
  --do_lower_case \
  --train_file $SQUAD_DIR/train-v2.0.json \
  --predict_file $SQUAD_DIR/dev-v2.0.json \
  --per_gpu_train_batch_size 3 \
  --per_gpu_eval_batch_size 64 \
  --learning_rate 3e-5 \
  --num_train_epochs 3.0 \
  --max_seq_length 384 \
  --doc_stride 128 \
  --save_steps 2000 \
  --threads 24 \
  --warmup_steps 814 \
  --gradient_accumulation_steps 4 \
  --fp16 \
  --do_train
Evaluation

Evaluation on the dev set. I did not sweep for best threshold.

val
exact 84.41842836688285
f1 87.4628460501696
total 11873.0
HasAns_exact 80.68488529014844
HasAns_f1 86.78245127423482
HasAns_total 5928.0
NoAns_exact 88.1412952060555
NoAns_f1 88.1412952060555
NoAns_total 5945.0
best_exact 84.41842836688285
best_exact_thresh 0.0
best_f1 87.46284605016956
best_f1_thresh 0.0
Usage

See huggingface documentation . Training on SQuAD V2 allows the model to score if a paragraph contains an answer: 

start_scores, end_scores = model(input_ids) 
span_scores = start_scores.softmax(dim=1).log()[:,:,None] + end_scores.softmax(dim=1).log()[:,None,:]
ignore_score = span_scores[:,0,0] #no answer scores

Runs of ktrapeznikov albert-xlarge-v2-squad-v2 on huggingface.co

109
Total runs
0
24-hour runs
16
3-day runs
30
7-day runs
-39
30-day runs

More Information About albert-xlarge-v2-squad-v2 huggingface.co Model

albert-xlarge-v2-squad-v2 huggingface.co

albert-xlarge-v2-squad-v2 huggingface.co is an AI model on huggingface.co that provides albert-xlarge-v2-squad-v2's model effect (), which can be used instantly with this ktrapeznikov albert-xlarge-v2-squad-v2 model. huggingface.co supports a free trial of the albert-xlarge-v2-squad-v2 model, and also provides paid use of the albert-xlarge-v2-squad-v2. Support call albert-xlarge-v2-squad-v2 model through api, including Node.js, Python, http.

albert-xlarge-v2-squad-v2 huggingface.co Url

https://huggingface.co/ktrapeznikov/albert-xlarge-v2-squad-v2

ktrapeznikov albert-xlarge-v2-squad-v2 online free

albert-xlarge-v2-squad-v2 huggingface.co is an online trial and call api platform, which integrates albert-xlarge-v2-squad-v2's modeling effects, including api services, and provides a free online trial of albert-xlarge-v2-squad-v2, you can try albert-xlarge-v2-squad-v2 online for free by clicking the link below.

ktrapeznikov albert-xlarge-v2-squad-v2 online free url in huggingface.co:

https://huggingface.co/ktrapeznikov/albert-xlarge-v2-squad-v2

albert-xlarge-v2-squad-v2 install

albert-xlarge-v2-squad-v2 is an open source model from GitHub that offers a free installation service, and any user can find albert-xlarge-v2-squad-v2 on GitHub to install. At the same time, huggingface.co provides the effect of albert-xlarge-v2-squad-v2 install, users can directly use albert-xlarge-v2-squad-v2 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

albert-xlarge-v2-squad-v2 install url in huggingface.co:

https://huggingface.co/ktrapeznikov/albert-xlarge-v2-squad-v2

Url of albert-xlarge-v2-squad-v2

albert-xlarge-v2-squad-v2 huggingface.co Url

Provider of albert-xlarge-v2-squad-v2 huggingface.co

ktrapeznikov
ORGANIZATIONS
huggingface.co

Other API from ktrapeznikov