metadata
library_name: transformers
pipeline_tag: text-generation
Direct Use
import transformers as tfm
model = tfm.AutoModelForCausalLM.from_pretrained("Owaner/fineweb-falcon")
tokenizer = tfm.PreTrainedTokenizerFast.from_pretrained("Owaner/falcon_tokenizer")
example = "When habitually indulge in "
tokenized_input = tokenizer(example, return_tensors="pt", return_token_type_ids=False)
output = model.generate(
inputs=tokenized_input["input_ids"],
attention_mask=tokenized_input["attention_mask"],
do_sample = True,
max_length=100,
temperature=0.7,
top_k=50,
top_p=0.95,
num_return_sequences=5
)
output_text = tokenizer.batch_decode(output, skip_special_tokens=True)
for i, o in enumerate(output_text):
print(f"Output {i+1}: {o}")
- Hardware Type: Single Nvidia A80 memory 80
- Hours used: 2 hours
- Cloud Provider: DataCrunch
- Compute Region: [More Information Needed]
- Carbon Emitted: [More Information Needed]