Update script usage examples to use filename + `classify` command.
Browse files
    	
        README.md
    CHANGED
    
    | @@ -22,7 +22,7 @@ This is a modified version of https://huggingface.co/datasets/uv-scripts/classif | |
| 22 |  | 
| 23 | 
             
            ```bash
         | 
| 24 | 
             
            # Classify IMDB reviews
         | 
| 25 | 
            -
            uv run  | 
| 26 | 
             
              --input-dataset stanfordnlp/imdb \
         | 
| 27 | 
             
              --column text \
         | 
| 28 | 
             
              --labels "positive,negative" \
         | 
| @@ -63,7 +63,7 @@ That's it! No installation, no setup - just `uv run`. | |
| 63 | 
             
            ### Basic Classification
         | 
| 64 |  | 
| 65 | 
             
            ```bash
         | 
| 66 | 
            -
            uv run  | 
| 67 | 
             
              --input-dataset <dataset-id> \
         | 
| 68 | 
             
              --column <text-column> \
         | 
| 69 | 
             
              --labels <comma-separated-labels> \
         | 
| @@ -98,7 +98,7 @@ uv run examples/classify-dataset.py \ | |
| 98 | 
             
            Provide context for your labels to improve classification accuracy:
         | 
| 99 |  | 
| 100 | 
             
            ```bash
         | 
| 101 | 
            -
            uv run  | 
| 102 | 
             
              --input-dataset user/support-tickets \
         | 
| 103 | 
             
              --column content \
         | 
| 104 | 
             
              --labels "bug,feature,question,other" \
         | 
| @@ -114,7 +114,7 @@ The model uses these descriptions to better understand what each label represent | |
| 114 | 
             
            Enable multi-label mode for documents that can have multiple applicable labels:
         | 
| 115 |  | 
| 116 | 
             
            ```bash
         | 
| 117 | 
            -
            uv run  | 
| 118 | 
             
              --input-dataset ag_news \
         | 
| 119 | 
             
              --column text \
         | 
| 120 | 
             
              --labels "world,sports,business,science" \
         | 
| @@ -128,7 +128,7 @@ uv run examples/classify-dataset.py \ | |
| 128 | 
             
            ### Sentiment Analysis
         | 
| 129 |  | 
| 130 | 
             
            ```bash
         | 
| 131 | 
            -
            uv run  | 
| 132 | 
             
              --input-dataset stanfordnlp/imdb \
         | 
| 133 | 
             
              --column text \
         | 
| 134 | 
             
              --labels "positive,ambivalent,negative" \
         | 
| @@ -139,7 +139,7 @@ uv run examples/classify-dataset.py \ | |
| 139 | 
             
            ### Support Ticket Classification
         | 
| 140 |  | 
| 141 | 
             
            ```bash
         | 
| 142 | 
            -
            uv run  | 
| 143 | 
             
              --input-dataset user/support-tickets \
         | 
| 144 | 
             
              --column content \
         | 
| 145 | 
             
              --labels "bug,feature_request,question,other" \
         | 
| @@ -151,7 +151,7 @@ uv run examples/classify-dataset.py \ | |
| 151 | 
             
            ### News Categorization
         | 
| 152 |  | 
| 153 | 
             
            ```bash
         | 
| 154 | 
            -
            uv run  | 
| 155 | 
             
              --input-dataset ag_news \
         | 
| 156 | 
             
              --column text \
         | 
| 157 | 
             
              --labels "world,sports,business,tech" \
         | 
| @@ -162,7 +162,7 @@ uv run examples/classify-dataset.py \ | |
| 162 | 
             
            ### Multi-Label News Classification
         | 
| 163 |  | 
| 164 | 
             
            ```bash
         | 
| 165 | 
            -
            uv run  | 
| 166 | 
             
              --input-dataset ag_news \
         | 
| 167 | 
             
              --column text \
         | 
| 168 | 
             
              --labels "world,sports,business,tech" \
         | 
| @@ -180,7 +180,7 @@ Classify academic papers into machine learning research areas: | |
| 180 |  | 
| 181 | 
             
            ```bash
         | 
| 182 | 
             
            # Fast classification with random sampling
         | 
| 183 | 
            -
            uv run  | 
| 184 | 
             
              --input-dataset librarian-bots/arxiv-metadata-snapshot \
         | 
| 185 | 
             
              --column abstract \
         | 
| 186 | 
             
              --labels "llm,computer_vision,reinforcement_learning,optimization,theory,other" \
         | 
| @@ -192,7 +192,7 @@ uv run examples/classify-dataset.py \ | |
| 192 | 
             
              --shuffle
         | 
| 193 |  | 
| 194 | 
             
            # Multi-label for nuanced classification
         | 
| 195 | 
            -
            uv run  | 
| 196 | 
             
              --input-dataset librarian-bots/arxiv-metadata-snapshot \
         | 
| 197 | 
             
              --column abstract \
         | 
| 198 | 
             
              --labels "multimodal,agents,reasoning,safety,efficiency" \
         | 
| @@ -212,7 +212,7 @@ This script is optimized to run locally on GPU-equipped machines: | |
| 212 |  | 
| 213 | 
             
            ```bash
         | 
| 214 | 
             
            # Local execution with your GPU
         | 
| 215 | 
            -
            uv run  | 
| 216 | 
             
              --input-dataset stanfordnlp/imdb \
         | 
| 217 | 
             
              --column text \
         | 
| 218 | 
             
              --labels "positive,negative" \
         | 
| @@ -230,7 +230,7 @@ When working with ordered datasets, use `--shuffle` with `--max-samples` to get | |
| 230 |  | 
| 231 | 
             
            ```bash
         | 
| 232 | 
             
            # Get 50 random reviews instead of the first 50
         | 
| 233 | 
            -
            uv run  | 
| 234 | 
             
              --input-dataset stanfordnlp/imdb \
         | 
| 235 | 
             
              --column text \
         | 
| 236 | 
             
              --labels "positive,negative" \
         | 
| @@ -248,7 +248,7 @@ By default, this script works with any instruction-tuned model. Here are some re | |
| 248 |  | 
| 249 | 
             
            ```bash
         | 
| 250 | 
             
            # Lightweight model for fast classification
         | 
| 251 | 
            -
            uv run  | 
| 252 | 
             
              --input-dataset user/my-dataset \
         | 
| 253 | 
             
              --column text \
         | 
| 254 | 
             
              --labels "A,B,C" \
         | 
| @@ -256,7 +256,7 @@ uv run examples/classify-dataset.py \ | |
| 256 | 
             
              --output-dataset user/classified
         | 
| 257 |  | 
| 258 | 
             
            # Larger model for complex classification
         | 
| 259 | 
            -
            uv run  | 
| 260 | 
             
              --input-dataset user/legal-docs \
         | 
| 261 | 
             
              --column text \
         | 
| 262 | 
             
              --labels "contract,patent,brief,memo,other" \
         | 
| @@ -264,7 +264,7 @@ uv run examples/classify-dataset.py \ | |
| 264 | 
             
              --output-dataset user/legal-classified
         | 
| 265 |  | 
| 266 | 
             
            # Specialized zero-shot classifier
         | 
| 267 | 
            -
            uv run  | 
| 268 | 
             
              --input-dataset user/my-dataset \
         | 
| 269 | 
             
              --column text \
         | 
| 270 | 
             
              --labels "A,B,C" \
         | 
| @@ -277,7 +277,7 @@ uv run examples/classify-dataset.py \ | |
| 277 | 
             
            Configure `--batch-size` for more effective batch processing with large datasets: 
         | 
| 278 |  | 
| 279 | 
             
            ```bash
         | 
| 280 | 
            -
            uv run  | 
| 281 | 
             
              --input-dataset user/huge-dataset \
         | 
| 282 | 
             
              --column text \
         | 
| 283 | 
             
              --labels "A,B,C" \
         | 
| @@ -342,7 +342,7 @@ Start with small tests, then run on the full dataset: | |
| 342 |  | 
| 343 | 
             
            ```bash
         | 
| 344 | 
             
            # Step 1: Test with small sample
         | 
| 345 | 
            -
            uv run  | 
| 346 | 
             
              --input-dataset your-dataset \
         | 
| 347 | 
             
              --column text \
         | 
| 348 | 
             
              --labels "label1,label2,label3" \
         | 
| @@ -351,7 +351,7 @@ uv run examples/classify-dataset.py \ | |
| 351 | 
             
              --max-samples 100
         | 
| 352 |  | 
| 353 | 
             
            # Step 2: If results look good, run on full dataset
         | 
| 354 | 
            -
            uv run  | 
| 355 | 
             
              --input-dataset your-dataset \
         | 
| 356 | 
             
              --column text \
         | 
| 357 | 
             
              --labels "label1,label2,label3" \
         | 
|  | |
| 22 |  | 
| 23 | 
             
            ```bash
         | 
| 24 | 
             
            # Classify IMDB reviews
         | 
| 25 | 
            +
            uv run classify-dataset.py classify \
         | 
| 26 | 
             
              --input-dataset stanfordnlp/imdb \
         | 
| 27 | 
             
              --column text \
         | 
| 28 | 
             
              --labels "positive,negative" \
         | 
|  | |
| 63 | 
             
            ### Basic Classification
         | 
| 64 |  | 
| 65 | 
             
            ```bash
         | 
| 66 | 
            +
            uv run classify-dataset.py classify \
         | 
| 67 | 
             
              --input-dataset <dataset-id> \
         | 
| 68 | 
             
              --column <text-column> \
         | 
| 69 | 
             
              --labels <comma-separated-labels> \
         | 
|  | |
| 98 | 
             
            Provide context for your labels to improve classification accuracy:
         | 
| 99 |  | 
| 100 | 
             
            ```bash
         | 
| 101 | 
            +
            uv run classify-dataset.py classify \
         | 
| 102 | 
             
              --input-dataset user/support-tickets \
         | 
| 103 | 
             
              --column content \
         | 
| 104 | 
             
              --labels "bug,feature,question,other" \
         | 
|  | |
| 114 | 
             
            Enable multi-label mode for documents that can have multiple applicable labels:
         | 
| 115 |  | 
| 116 | 
             
            ```bash
         | 
| 117 | 
            +
            uv run classify-dataset.py classify \
         | 
| 118 | 
             
              --input-dataset ag_news \
         | 
| 119 | 
             
              --column text \
         | 
| 120 | 
             
              --labels "world,sports,business,science" \
         | 
|  | |
| 128 | 
             
            ### Sentiment Analysis
         | 
| 129 |  | 
| 130 | 
             
            ```bash
         | 
| 131 | 
            +
            uv run classify-dataset.py classify \
         | 
| 132 | 
             
              --input-dataset stanfordnlp/imdb \
         | 
| 133 | 
             
              --column text \
         | 
| 134 | 
             
              --labels "positive,ambivalent,negative" \
         | 
|  | |
| 139 | 
             
            ### Support Ticket Classification
         | 
| 140 |  | 
| 141 | 
             
            ```bash
         | 
| 142 | 
            +
            uv run classify-dataset.py classify \
         | 
| 143 | 
             
              --input-dataset user/support-tickets \
         | 
| 144 | 
             
              --column content \
         | 
| 145 | 
             
              --labels "bug,feature_request,question,other" \
         | 
|  | |
| 151 | 
             
            ### News Categorization
         | 
| 152 |  | 
| 153 | 
             
            ```bash
         | 
| 154 | 
            +
            uv run classify-dataset.py classify \
         | 
| 155 | 
             
              --input-dataset ag_news \
         | 
| 156 | 
             
              --column text \
         | 
| 157 | 
             
              --labels "world,sports,business,tech" \
         | 
|  | |
| 162 | 
             
            ### Multi-Label News Classification
         | 
| 163 |  | 
| 164 | 
             
            ```bash
         | 
| 165 | 
            +
            uv run classify-dataset.py classify \
         | 
| 166 | 
             
              --input-dataset ag_news \
         | 
| 167 | 
             
              --column text \
         | 
| 168 | 
             
              --labels "world,sports,business,tech" \
         | 
|  | |
| 180 |  | 
| 181 | 
             
            ```bash
         | 
| 182 | 
             
            # Fast classification with random sampling
         | 
| 183 | 
            +
            uv run classify-dataset.py classify \
         | 
| 184 | 
             
              --input-dataset librarian-bots/arxiv-metadata-snapshot \
         | 
| 185 | 
             
              --column abstract \
         | 
| 186 | 
             
              --labels "llm,computer_vision,reinforcement_learning,optimization,theory,other" \
         | 
|  | |
| 192 | 
             
              --shuffle
         | 
| 193 |  | 
| 194 | 
             
            # Multi-label for nuanced classification
         | 
| 195 | 
            +
            uv run classify-dataset.py classify \
         | 
| 196 | 
             
              --input-dataset librarian-bots/arxiv-metadata-snapshot \
         | 
| 197 | 
             
              --column abstract \
         | 
| 198 | 
             
              --labels "multimodal,agents,reasoning,safety,efficiency" \
         | 
|  | |
| 212 |  | 
| 213 | 
             
            ```bash
         | 
| 214 | 
             
            # Local execution with your GPU
         | 
| 215 | 
            +
            uv run classify-dataset.py classify \
         | 
| 216 | 
             
              --input-dataset stanfordnlp/imdb \
         | 
| 217 | 
             
              --column text \
         | 
| 218 | 
             
              --labels "positive,negative" \
         | 
|  | |
| 230 |  | 
| 231 | 
             
            ```bash
         | 
| 232 | 
             
            # Get 50 random reviews instead of the first 50
         | 
| 233 | 
            +
            uv run classify-dataset.py classify \
         | 
| 234 | 
             
              --input-dataset stanfordnlp/imdb \
         | 
| 235 | 
             
              --column text \
         | 
| 236 | 
             
              --labels "positive,negative" \
         | 
|  | |
| 248 |  | 
| 249 | 
             
            ```bash
         | 
| 250 | 
             
            # Lightweight model for fast classification
         | 
| 251 | 
            +
            uv run classify-dataset.py classify \
         | 
| 252 | 
             
              --input-dataset user/my-dataset \
         | 
| 253 | 
             
              --column text \
         | 
| 254 | 
             
              --labels "A,B,C" \
         | 
|  | |
| 256 | 
             
              --output-dataset user/classified
         | 
| 257 |  | 
| 258 | 
             
            # Larger model for complex classification
         | 
| 259 | 
            +
            uv run classify-dataset.py classify \
         | 
| 260 | 
             
              --input-dataset user/legal-docs \
         | 
| 261 | 
             
              --column text \
         | 
| 262 | 
             
              --labels "contract,patent,brief,memo,other" \
         | 
|  | |
| 264 | 
             
              --output-dataset user/legal-classified
         | 
| 265 |  | 
| 266 | 
             
            # Specialized zero-shot classifier
         | 
| 267 | 
            +
            uv run classify-dataset.py classify \
         | 
| 268 | 
             
              --input-dataset user/my-dataset \
         | 
| 269 | 
             
              --column text \
         | 
| 270 | 
             
              --labels "A,B,C" \
         | 
|  | |
| 277 | 
             
            Configure `--batch-size` for more effective batch processing with large datasets: 
         | 
| 278 |  | 
| 279 | 
             
            ```bash
         | 
| 280 | 
            +
            uv run classify-dataset.py classify \
         | 
| 281 | 
             
              --input-dataset user/huge-dataset \
         | 
| 282 | 
             
              --column text \
         | 
| 283 | 
             
              --labels "A,B,C" \
         | 
|  | |
| 342 |  | 
| 343 | 
             
            ```bash
         | 
| 344 | 
             
            # Step 1: Test with small sample
         | 
| 345 | 
            +
            uv run classify-dataset.py classify \
         | 
| 346 | 
             
              --input-dataset your-dataset \
         | 
| 347 | 
             
              --column text \
         | 
| 348 | 
             
              --labels "label1,label2,label3" \
         | 
|  | |
| 351 | 
             
              --max-samples 100
         | 
| 352 |  | 
| 353 | 
             
            # Step 2: If results look good, run on full dataset
         | 
| 354 | 
            +
            uv run classify-dataset.py classify \
         | 
| 355 | 
             
              --input-dataset your-dataset \
         | 
| 356 | 
             
              --column text \
         | 
| 357 | 
             
              --labels "label1,label2,label3" \
         | 
