modifying name of the metric

README.md

@@ -1,5 +1,5 @@
 ---
-title: Perplexity
+title: DmxPerplexity
 emoji: 🌖
 colorFrom: purple
 colorTo: pink
@@ -29,19 +29,19 @@ It is defined as the exponentiated average negative log-likelihood of a sequence
 For more information, see https://huggingface.co/docs/transformers/perplexity
 
 ## How to Use
-At minimum, this metric requires the model and data as inputs.
+At minimum, this metric requires the model and references as inputs.
 ```python
 >>> import evaluate
->>> perplexity = evaluate.load("perplexity", module_type="metric")
+>>> perplexity = evaluate.load("dmx_perplexity", module_type="metric")
 >>> input_texts = ["lorem ipsum", "Happy Birthday!", "Bienvenue"]
->>> results = perplexity.compute(model='distilgpt2',data=input_texts)
+>>> results = perplexity.compute(model='distilgpt2',references=input_texts)
 >>> print(results)
-{'accuracy': 1.0}
+{'loss': 4.993086338043213, 'perplexity': 147.390625}
 ```
 
 ### Inputs
 - **model** (`Union`[`str`,`AutoModelForCausalLM`]): model used for calculating Perplexity
-- **data** (`list` of `str`): input text, each separate text snippet is one list entry.
+- **references** (`list` of `str`): input text, each separate text snippet is one list entry.
 - **device** (`str`): device to run on, defaults to 'cuda' when available.
 - **max_length** (`int`): maximum sequence length, defaults to 2048.
 
@@ -59,9 +59,9 @@ This metric outputs a dictionary, containing the loss and perplexity score.
 ```python
 >>> import evaluate
 >>> from datasets import load_dataset
->>> perplexity = evaluate.load("d-matrix/perplexity", module_type="metric")
+>>> perplexity = evaluate.load("dmx_perplexity", module_type="metric")
 >>> input_texts = load_dataset("wikitext", "wikitext-2-raw-v1", split="test")["text"][:10] 
->>> results = perplexity.compute(model='distilgpt2',data=input_texts)
+>>> results = perplexity.compute(model='distilgpt2',references=input_texts)
 >>> print(list(results.keys()))
 ['loss', 'perplexity']
 >>> print(results['loss']) 

perplexity.py → dmx_perplexity/dmx_perplexity.py

@@ -20,7 +20,7 @@ Args:
                     causal versions of t5, and more (the full list can be found
                     in the AutoModelForCausalLM documentation here:
                     https://huggingface.co/docs/transformers/master/en/model_doc/auto#transformers.AutoModelForCausalLM )
-    data (list of str): input text, each separate text snippet is one list entry.
+    references (list of str): input text, each separate text snippet is one list entry.
     device (str): device to run on, defaults to 'cuda' when available.
     max_length (int): maximum sequence length, defaults to 2048.
 Returns:
@@ -31,7 +31,7 @@ Examples:
         >>> perplexity = evaluate.load("dmx_perplexity", module_type="metric")
         >>> input_texts = load_dataset("wikitext", "wikitext-2-raw-v1", split="test")["text"][:10] # doctest: +SKIP
         >>> results = perplexity.compute(model='distilgpt2',
-        ...                              data=input_texts)
+        ...                              references=input_texts)
         >>> print(list(results.keys()))
         ['loss', 'perplexity']
         >>> print(results['loss']) # doctest: +SKIP
@@ -41,7 +41,7 @@ Examples:
 """
 
 @evaluate.utils.file_utils.add_start_docstrings(_DESCRIPTION, _KWARGS_DESCRIPTION)
-class Perplexity(evaluate.Metric):
+class DmxPerplexity(evaluate.Metric):
     def _info(self):
         return evaluate.MetricInfo(
             module_type="metric",
@@ -50,7 +50,7 @@ class Perplexity(evaluate.Metric):
             inputs_description=_KWARGS_DESCRIPTION,
             features=datasets.Features(
                 {
-                    "text": datasets.Value("string"),
+                    "references": datasets.Value("string"),
                 }
             ),
             reference_urls=["https://huggingface.co/docs/transformers/perplexity"],
@@ -58,7 +58,7 @@ class Perplexity(evaluate.Metric):
 
     def _compute(
         self,
-        data,
+        references,
         model: Union[str, AutoModelForCausalLM],
         device=None,
         max_length=None,
@@ -91,7 +91,7 @@ class Perplexity(evaluate.Metric):
             max_seq_len = 2048
 
         model = model.to(device)
-        encodings = tokenizer("\n\n".join(data), return_tensors="pt")
+        encodings = tokenizer("\n\n".join(references), return_tensors="pt")
 
         stride = max_seq_len
         seq_len = encodings.input_ids.size(1)