Update README.md

README.md

@@ -53,7 +53,31 @@ Given the training format, no extra care is needed to account for different sequ
 
 ## Performance
 The resulting model matches SOTA performance with 82.5% accuracy.
+```
+Total number of examples                 9832
+Number with badly formed result          0
+Number with incorrect result             1725
+Number with correct result               8107 
+[82.5%]
+
+example 0 = {'prompt_text': "mnli hypothesis: Your contributions were of no help with our students' education.
+                            premise: Your contribution helped make it possible for us to provide our students
+                            with a quality education. target:",
+             'class_label': 'contradiction'}
+result = {'generated_text': ' contradiction'}
 
+First 10 generated_text and expected class_label results:
+ 0: 'contradiction'                          contradiction
+ 1: 'contradiction'                          contradiction
+ 2: 'entailment'                             entailment
+ 3: 'contradiction'                          contradiction
+ 4: 'entailment'                             entailment
+ 5: 'entailment'                             entailment
+ 6: 'contradiction'                          contradiction
+ 7: 'contradiction'                          contradiction
+ 8: 'entailment'                             neutral
+ 9: 'contradiction'                          contradiction
+```
 ## How to use
 The model can be easily loaded using AutoModelForCausalLM.
 You can use the pipeline API for text generation.
@@ -61,8 +85,8 @@ You can use the pipeline API for text generation.
 ```python
 from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer
 
-hf_model = AutoModelForCausalLM.from_pretrained("Graphcore/gptj-mnli")
 tokenizer = AutoTokenizer.from_pretrained('EleutherAI/gpt-j-6B')
+hf_model = AutoModelForCausalLM.from_pretrained("Graphcore/gptj-mnli", pad_token_id=tokenizer.eos_token_id)
 generator =  pipeline('text-generation', model=hf_model, tokenizer=tokenizer)
 prompt = "mnli hypothesis: Your contributions were of no help with our students' education." \
          "premise: Your contribution helped make it possible for us to provide our students with a quality education. target:"