Edit model card

To reproduce this run, execute:

#!/usr/bin/env bash
python run_flax_speech_recognition_seq2seq.py \
        --dataset_name="esb/datasets" \
        --model_name_or_path="esb/wav2vec2-aed-tedlium" \
        --dataset_config_name="tedlium" \
        --output_dir="./" \
        --wandb_name="wav2vec2-aed-tedlium" \
        --wandb_project="wav2vec2-aed" \
        --per_device_train_batch_size="8" \
        --per_device_eval_batch_size="2" \
        --learning_rate="1e-4" \
        --warmup_steps="500" \
        --logging_steps="25" \
        --max_steps="50001" \
        --eval_steps="10000" \
        --save_steps="10000" \
        --generation_max_length="40" \
        --generation_num_beams="1" \
        --final_generation_max_length="250" \
        --final_generation_num_beams="12" \
        --generation_length_penalty="1.5" \
        --hidden_dropout="0.2" \
        --activation_dropout="0.2" \
        --feat_proj_dropout="0.2" \
        --overwrite_output_dir \
        --gradient_checkpointing \
        --freeze_feature_encoder \
        --predict_with_generate \
        --do_eval \
        --do_train \
        --do_predict \
        --push_to_hub \
        --use_auth_token
Downloads last month
3
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Datasets used to train esc-bench/wav2vec2-aed-tedlium