shorecode/t5-efficient-tiny-summarizer-general-purpose-v3

1---2license: mit3datasets:4- shorecode/summary-collection-200k-rows5language:6- en7base_model:8- google/t5-efficient-tiny-nh89library_name: transformers10tags:11- summary12- summarizer13widget:14- text: Model training15  output:16    url: Screenshot_20251104_204645.png17metrics:18- f119- rouge20- extractiveness21model-index:22- name: t5-efficient-tiny-summarizer-general-purpose-v223  results:24  - task:25      type: Summarization26    dataset:27      name: shorecode/summary-collection-60k-rows28      type: shorecode/summary-collection-60k-rows29    metrics:30    - name: f1 Score31      type: f1 Score32      value: 0.2933  - task:34      type: Summarization35    dataset:36      name: shorecode/summary-collection-60k-rows37      type: shorecode/summary-collection-60k-rows38    metrics:39    - name: Faithfullness (facebook/bart-large-cnn)40      type: facebook/bart-large-cnn41      value: 1.7142  - task:43      type: Summarization44    dataset:45      name: shorecode/summary-collection-60k-rows46      type: shorecode/summary-collection-60k-rows47    metrics:48    - name: Summarization Compression49      type: Lighteval extractiveness50      value: 7.5251  - task:52      type: Summarization53    dataset:54      name: shorecode/summary-collection-60k-rows55      type: shorecode/summary-collection-60k-rows56    metrics:57    - name: Summarization Coverage58      type: Lighteval extractiveness59      value: 0.9660  - task:61      type: Summarization62    dataset:63      name: shorecode/summary-collection-60k-rows64      type: shorecode/summary-collection-60k-rows65    metrics:66    - name: Summarization Density67      type: Lighteval extractiveness68      value: 8.6869  - task:70      type: Summarization71    dataset:72      name: shorecode/summary-collection-60k-rows73      type: shorecode/summary-collection-60k-rows74    metrics:75    - name: rougeL precision76      type: Lighteval77      value: 0.5978  - task:79      type: Summarization80    dataset:81      name: shorecode/summary-collection-60k-rows82      type: shorecode/summary-collection-60k-rows83    metrics:84    - name: rougeL recall85      type: Lighteval86      value: 0.3187  - task:88      type: Summarization89    dataset:90      name: shorecode/summary-collection-60k-rows91      type: shorecode/summary-collection-60k-rows92    metrics:93    - name: rougeL fmeasure94      type: Lighteval95      value: 0.4196  - task:97      type: Summarization98    dataset:99      name: shorecode/summary-collection-60k-rows100      type: shorecode/summary-collection-60k-rows101    metrics:102    - name: rouge1 precision103      type: Lighteval104      value: 0.63105  - task:106      type: Summarization107    dataset:108      name: shorecode/summary-collection-60k-rows109      type: shorecode/summary-collection-60k-rows110    metrics:111    - name: rouge1 recall112      type: Lighteval113      value: 0.33114  - task:115      type: Summarization116    dataset:117      name: shorecode/summary-collection-60k-rows118      type: shorecode/summary-collection-60k-rows119    metrics:120    - name: rouge1 fmeasure121      type: Lighteval122      value: 0.44123---124 125# This model was built to shorten text that is injected into LLM prompts to reduce API calling costs126Very high compression (7x+) meaning the text is 7 times smaller when sent to your LLM provider!127Recommended kwargs:128- num_beams=2 || 3129- no_repeat_ngram_size=2130- min_length=20131- max_new_tokens=500 || as high as you can tolerate, 4x500 = 2000 characters. anything after 2000 is clipped132<Gallery />133 134https://api.wandb.ai/links/shorecode-shorecode-llc/6udfudmr