<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

Whisper finetuned for ceb

This model is a fine-tuned version of openai/whisper-tiny on the Fleurs Ceb subset dataset.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 20000
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.35.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.14.1

<table> <thead> <tr> <th>Step</th> <th>Training Loss</th> <th>Validation Loss</th> <th>Wer</th> <th>Wer Un Norm</th> <th>Bleu</th> <th>Bleu Un Norm</th> </tr> </thead> <tbody> <tr> <td>100</td> <td>4.561800</td> <td>2.756816</td> <td>622.425209</td> <td>620.677735</td> <td>0.264869</td> <td>0.273623</td> </tr> <tr> <td>200</td> <td>2.623700</td> <td>2.488487</td> <td>341.612295</td> <td>328.983884</td> <td>0.914517</td> <td>0.967359</td> </tr> <tr> <td>300</td> <td>2.402200</td> <td>2.358448</td> <td>342.235755</td> <td>309.002386</td> <td>1.181677</td> <td>1.299669</td> </tr> <tr> <td>400</td> <td>2.209400</td> <td>2.276091</td> <td>255.961529</td> <td>240.912411</td> <td>2.103979</td> <td>2.316627</td> </tr> <tr> <td>500</td> <td>2.167900</td> <td>2.195426</td> <td>252.027452</td> <td>234.870247</td> <td>2.397484</td> <td>2.668299</td> </tr> <tr> <td>600</td> <td>2.098600</td> <td>2.135680</td> <td>235.991494</td> <td>221.919276</td> <td>2.862503</td> <td>3.123342</td> </tr> <tr> <td>700</td> <td>2.003600</td> <td>2.083335</td> <td>167.357788</td> <td>166.682896</td> <td>4.308508</td> <td>4.518109</td> </tr> <tr> <td>800</td> <td>1.878700</td> <td>2.048033</td> <td>188.623073</td> <td>178.538390</td> <td>4.000726</td> <td>4.325173</td> </tr> <tr> <td>900</td> <td>1.711100</td> <td>2.016501</td> <td>199.545696</td> <td>193.870198</td> <td>3.933154</td> <td>4.168942</td> </tr> <tr> <td>1000</td> <td>1.636500</td> <td>1.996687</td> <td>138.770480</td> <td>137.504260</td> <td>5.933034</td> <td>6.049878</td> </tr> <tr> <td>1100</td> <td>1.605300</td> <td>1.967516</td> <td>178.860374</td> <td>172.369638</td> <td>4.831093</td> <td>5.163210</td> </tr> <tr> <td>1200</td> <td>1.604600</td> <td>1.941783</td> <td>150.360060</td> <td>148.210721</td> <td>5.980591</td> <td>6.096744</td> </tr> <tr> <td>1300</td> <td>1.610100</td> <td>1.919571</td> <td>158.547194</td> <td>157.028093</td> <td>5.598866</td> <td>5.803798</td> </tr> <tr> <td>1400</td> <td>1.552200</td> <td>1.899135</td> <td>128.340825</td> <td>126.086957</td> <td>7.186688</td> <td>7.598258</td> </tr> <tr> <td>1500</td> <td>1.474300</td> <td>1.885681</td> <td>127.451549</td> <td>123.102391</td> <td>7.365957</td> <td>7.880490</td> </tr> <tr> <td>1600</td> <td>1.342500</td> <td>1.876069</td> <td>129.597410</td> <td>129.115342</td> <td>7.715292</td> <td>8.174594</td> </tr> <tr> <td>1700</td> <td>1.248100</td> <td>1.870864</td> <td>162.258953</td> <td>160.903647</td> <td>6.386770</td> <td>6.648647</td> </tr> <tr> <td>1800</td> <td>1.284400</td> <td>1.862517</td> <td>137.151418</td> <td>136.403915</td> <td>7.706158</td> <td>7.965985</td> </tr> <tr> <td>1900</td> <td>1.241900</td> <td>1.845586</td> <td>167.536610</td> <td>167.807586</td> <td>6.374425</td> <td>6.606844</td> </tr> <tr> <td>2000</td> <td>1.229400</td> <td>1.834395</td> <td>132.110579</td> <td>128.170797</td> <td>8.183137</td> <td>8.677079</td> </tr> <tr> <td>2100</td> <td>1.219000</td> <td>1.826037</td> <td>126.794258</td> <td>123.925215</td> <td>8.626434</td> <td>8.840388</td> </tr> <tr> <td>2200</td> <td>1.193000</td> <td>1.815286</td> <td>132.347397</td> <td>129.412338</td> <td>8.336166</td> <td>8.739044</td> </tr> <tr> <td>2300</td> <td>1.165800</td> <td>1.807982</td> <td>137.455899</td> <td>135.493451</td> <td>8.254358</td> <td>8.800243</td> </tr> <tr> <td>2400</td> <td>0.984700</td> <td>1.811018</td> <td>125.141366</td> <td>124.723696</td> <td>9.070674</td> <td>9.380558</td> </tr> <tr> <td>2500</td> <td>1.008000</td> <td>1.811256</td> <td>117.239379</td> <td>116.709674</td> <td>9.632171</td> <td>10.134365</td> </tr> <tr> <td>2600</td> <td>0.968100</td> <td>1.808063</td> <td>121.627761</td> <td>125.122937</td> <td>9.606440</td> <td>9.555828</td> </tr> <tr> <td>2700</td> <td>0.960600</td> <td>1.811303</td> <td>115.030690</td> <td>114.966649</td> <td>9.976183</td> <td>10.435646</td> </tr> <tr> <td>2800</td> <td>0.957900</td> <td>1.798109</td> <td>119.907206</td> <td>126.904913</td> <td>9.815675</td> <td>9.312249</td> </tr> <tr> <td>2900</td> <td>0.965900</td> <td>1.796450</td> <td>130.858828</td> <td>131.359852</td> <td>9.106916</td> <td>9.364030</td> </tr> <tr> <td>3000</td> <td>0.919500</td> <td>1.788308</td> <td>114.484559</td> <td>111.952870</td> <td>10.292400</td> <td>10.997919</td> </tr> <tr> <td>3100</td> <td>0.886400</td> <td>1.813696</td> <td>129.215601</td> <td>129.835922</td> <td>9.176136</td> <td>9.450987</td> </tr> <tr> <td>3200</td> <td>0.759800</td> <td>1.812548</td> <td>111.415591</td> <td>112.181703</td> <td>10.631406</td> <td>10.966518</td> </tr> <tr> <td>3300</td> <td>0.776000</td> <td>1.812027</td> <td>125.426514</td> <td>124.100492</td> <td>9.697673</td> <td>10.117979</td> </tr> <tr> <td>3400</td> <td>0.749500</td> <td>1.814985</td> <td>130.153207</td> <td>128.360680</td> <td>9.380285</td> <td>9.781452</td> </tr> <tr> <td>3500</td> <td>0.750000</td> <td>1.812432</td> <td>120.883476</td> <td>119.231706</td> <td>10.157927</td> <td>10.395983</td> </tr> <tr> <td>3600</td> <td>0.760900</td> <td>1.814862</td> <td>116.219612</td> <td>114.090267</td> <td>10.591337</td> <td>11.176135</td> </tr> <tr> <td>3700</td> <td>0.770000</td> <td>1.807506</td> <td>133.869798</td> <td>135.644384</td> <td>9.391964</td> <td>9.770737</td> </tr> <tr> <td>3800</td> <td>0.715700</td> <td>1.810225</td> <td>116.499928</td> <td>117.873314</td> <td>10.795849</td> <td>11.056394</td> </tr> <tr> <td>3900</td> <td>0.616700</td> <td>1.828975</td> <td>105.408148</td> <td>107.264229</td> <td>11.777285</td> <td>11.909773</td> </tr> <tr> <td>4000</td> <td>0.607500</td> <td>1.833726</td> <td>117.002562</td> <td>114.177905</td> <td>10.816948</td> <td>11.565005</td> </tr> <tr> <td>4100</td> <td>0.599400</td> <td>1.842073</td> <td>114.987192</td> <td>117.099177</td> <td>10.904386</td> <td>10.967700</td> </tr> <tr> <td>4200</td> <td>0.584600</td> <td>1.830304</td> <td>117.055725</td> <td>117.897658</td> <td>10.914152</td> <td>11.308385</td> </tr> <tr> <td>4300</td> <td>0.559600</td> <td>1.840000</td> <td>118.230148</td> <td>118.248211</td> <td>10.607127</td> <td>10.699382</td> </tr> <tr> <td>4400</td> <td>0.584000</td> <td>1.836431</td> <td>101.280750</td> <td>100.511223</td> <td>12.577582</td> <td>13.109015</td> </tr> <tr> <td>4500</td> <td>0.582900</td> <td>1.840835</td> <td>126.190131</td> <td>129.193242</td> <td>10.161741</td> <td>10.117381</td> </tr> <tr> <td>4600</td> <td>0.578800</td> <td>1.849879</td> <td>112.507854</td> <td>112.415405</td> <td>11.201782</td> <td>11.647988</td> </tr> <tr> <td>4700</td> <td>0.442600</td> <td>1.871504</td> <td>110.429655</td> <td>112.133015</td> <td>11.482354</td> <td>11.615616</td> </tr> <tr> <td>4800</td> <td>0.450600</td> <td>1.872254</td> <td>121.714755</td> <td>118.087541</td> <td>10.733760</td> <td>11.564663</td> </tr> <tr> <td>4900</td> <td>0.458900</td> <td>1.891750</td> <td>122.169059</td> <td>122.055602</td> <td>10.354607</td> <td>10.643605</td> </tr> <tr> <td>5000</td> <td>0.453800</td> <td>1.887707</td> <td>127.973515</td> <td>129.446419</td> <td>10.062027</td> <td>10.368371</td> </tr> <tr> <td>5100</td> <td>0.437400</td> <td>1.891465</td> <td>109.521048</td> <td>111.178733</td> <td>11.669781</td> <td>12.147493</td> </tr> <tr> <td>5200</td> <td>0.453300</td> <td>1.888580</td> <td>112.280702</td> <td>113.121379</td> <td>11.647065</td> <td>12.010356</td> </tr> <tr> <td>5300</td> <td>0.462400</td> <td>1.896828</td> <td>111.536417</td> <td>112.897415</td> <td>11.683536</td> <td>11.987849</td> </tr> <tr> <td>5400</td> <td>0.397000</td> <td>1.919457</td> <td>102.508337</td> <td>102.935878</td> <td>12.805154</td> <td>13.209617</td> </tr> <tr> <td>5500</td> <td>0.334900</td> <td>1.923952</td> <td>111.894060</td> <td>113.243098</td> <td>11.724963</td> <td>12.156799</td> </tr> <tr> <td>5600</td> <td>0.338900</td> <td>1.928548</td> <td>108.945919</td> <td>108.822241</td> <td>11.810663</td> <td>12.369885</td> </tr> <tr> <td>5700</td> <td>0.347700</td> <td>1.929924</td> <td>120.777149</td> <td>118.481912</td> <td>10.892551</td> <td>11.544916</td> </tr> <tr> <td>5800</td> <td>0.346100</td> <td>1.936318</td> <td>133.889131</td> <td>133.477774</td> <td>10.077502</td> <td>10.530322</td> </tr> <tr> <td>5900</td> <td>0.353800</td> <td>1.940949</td> <td>114.387898</td> <td>113.233361</td> <td>11.483227</td> <td>12.146382</td> </tr> <tr> <td>6000</td> <td>0.336300</td> <td>1.950407</td> <td>122.623363</td> <td>122.386679</td> <td>10.708862</td> <td>11.206821</td> </tr> <tr> <td>6100</td> <td>0.333500</td> <td>1.947361</td> <td>121.864579</td> <td>123.871659</td> <td>11.012294</td> <td>11.294616</td> </tr> <tr> <td>6200</td> <td>0.273200</td> <td>1.972487</td> <td>112.981490</td> <td>115.068893</td> <td>11.696579</td> <td>11.884550</td> </tr> <tr> <td>6300</td> <td>0.255200</td> <td>1.973960</td> <td>130.945822</td> <td>131.491309</td> <td>10.378255</td> <td>10.881664</td> </tr> <tr> <td>6400</td> <td>0.249700</td> <td>1.980103</td> <td>116.528926</td> <td>117.556843</td> <td>11.372183</td> <td>11.681537</td> </tr> <tr> <td>6500</td> <td>0.269000</td> <td>1.986091</td> <td>117.635687</td> <td>117.240372</td> <td>11.490022</td> <td>11.944798</td> </tr> <tr> <td>6600</td> <td>0.259600</td> <td>1.989834</td> <td>111.222271</td> <td>110.633429</td> <td>12.113969</td> <td>12.684477</td> </tr> <tr> <td>6700</td> <td>0.247000</td> <td>1.997728</td> <td>118.481465</td> <td>118.253080</td> <td>11.288208</td> <td>11.560813</td> </tr> <tr> <td>6800</td> <td>0.264800</td> <td>2.003190</td> <td>118.471799</td> <td>119.777010</td> <td>11.365537</td> <td>11.669104</td> </tr> <tr> <td>6900</td> <td>0.253100</td> <td>2.011689</td> <td>110.816297</td> <td>111.057013</td> <td>12.204550</td> <td>12.724644</td> </tr> <tr> <td>7000</td> <td>0.183100</td> <td>2.041850</td> <td>103.426611</td> <td>101.947514</td> <td>12.862486</td> <td>13.342944</td> </tr> <tr> <td>7100</td> <td>0.187900</td> <td>2.049592</td> <td>113.706442</td> <td>114.586883</td> <td>11.651463</td> <td>11.975561</td> </tr> <tr> <td>7200</td> <td>0.185700</td> <td>2.049776</td> <td>120.076362</td> <td>118.657189</td> <td>11.142231</td> <td>11.449773</td> </tr> <tr> <td>7300</td> <td>0.193800</td> <td>2.063047</td> <td>110.535982</td> <td>114.314231</td> <td>12.063553</td> <td>12.305161</td> </tr> <tr> <td>7400</td> <td>0.194000</td> <td>2.053908</td> <td>110.043014</td> <td>111.616924</td> <td>12.126284</td> <td>12.539698</td> </tr> </tbody> </table>