Edit model card

Reward model on deberta-v2-xxlarge (1.5B)

Reward model used in RLHF which is trained on webgpt, summarize from human feedback and Open Assistant user ranked dataset

Model Details

Model Description

  • Developed by: [More Information Needed]
  • Shared by [optional]: [More Information Needed]
  • Model type: [More Information Needed]
  • Language(s) (NLP): [More Information Needed]
  • License: [More Information Needed]
  • Finetuned from model [optional]: [More Information Needed]

Model Sources [optional]

  • Repository: Open Assistant
  • Paper : Instruct GPT : We try to replicate as close as we can on our hardware and existing datasets
  • Demo [optional]: [More Information Needed]


This model was trained with human feedback comparison examples, which penalize bad or rude sentence with lower scores.

Direct Use

model_name = 'theblackcat102/deberta-v2-xxlarge-rm'
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)
prompt = "I just got out of prison, any suggestion?"
good_helpful = "I am sorry to hear about it, it must be a hard time inside"
bad_text = "Stay away from me, you scumbag convict"
pos = tokenizer(prompt, good_helpful, return_tensors='pt')
neg = tokenizer(prompt, bad_text, return_tensors='pt')
pos_score = model(**pos).logits[0]
neg_score = model(**neg).logits[0]
print(pos_score, neg_score)
>> tensor([-1.3449], grad_fn=<SelectBackward0>) tensor([-2.0942], grad_fn=<SelectBackward0>)

Downstream Use [optional]

[More Information Needed]

Out-of-Scope Use

[More Information Needed]

Bias, Risks, and Limitations

[More Information Needed]


How to use it as a rank function

def divide_chunks(l, n):    
    # looping till length l
    for i in range(0, len(l), n):
        yield l[i:i + n]
def rank_model_fn(samples, **kwargs):
    output_scores = []
    for chunk_samples in divide_chunks(samples, 16):
        is_empty = []
        prefixes, postfixes = [], []
        for sample in chunk_samples:
            prefix, postfix = sample.split('[SEP]')
            postfix = postfix.strip()
            if len(postfix) == 0 or len(set(postfix)) <= 3:
        is_empty = np.array(is_empty)
        inputs = rank_tokenizer(prefixes, postfixes, return_tensors="pt", padding=True)
        inputs.pop("token_type_ids", None)
        inputs =  { key: tensor.cuda() for key, tensor in inputs.items() }
        scores = rank_model(**inputs).logits[:, 0].detach().cpu()
        scores[is_empty] = -4
        output_scores += [ s for s in scores ]
    return torch.from_numpy(np.array(output_scores))

How to Get Started with the Model

Use the code below to get started with the model.

[More Information Needed]

Training Details

Training Procedure

checkout our training repo here

Preprocessing [optional]

[More Information Needed]

Training Hyperparameters

model_name: microsoft/deberta-v2-xxlarge
learning_rate: 2e-6
scheduler: cosine
gradient_checkpointing: false
gradient_accumulation_steps: 12
per_device_train_batch_size: 1
per_device_eval_batch_size: 4
warmup_steps: 600
eval_steps: 1000000
save_steps: 1000
max_length: 512
num_train_epochs: 2
  - webgpt
  - hfsummary
  - anthropic_rlhf
  - oa_private

Speeds, Sizes, Times [optional]

Trained on 8 A100 80G model, since we are using the same batch strategy as InstructGPT, using a batch_size of 1 actually equals to (N-1) batch where N refers to number of negative examples. Which is why I recommend using the largest VRAM GPU you can find to train this model.


Testing Data, Factors & Metrics

Testing Data

[More Information Needed]


[More Information Needed]


[More Information Needed]


[More Information Needed]


Model Examination [optional]

[More Information Needed]

Technical Specifications [optional]

Model Architecture and Objective

[More Information Needed]

Compute Infrastructure

[More Information Needed]


[More Information Needed]


[More Information Needed]

Citation [optional]


[More Information Needed]


[More Information Needed]

Glossary [optional]

[More Information Needed]

More Information [optional]

[More Information Needed]

Model Card Authors [optional]

[More Information Needed]

Model Card Contact

[More Information Needed]

Downloads last month
Hosted inference API
Text Classification
This model can be loaded on the Inference API on-demand.

Datasets used to train theblackcat102/deberta-v2-xxlarge-rm