NeuralStockFusion-7b

merge

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the Model Stock merge method using Kukedlc/NeuralSirKrishna-7b as a base.

Models Merged

The following models were included in the merge:

• Kukedlc/NeuralArjuna-7B-DT
• Kukedlc/NeuralMaths-Experiment-7b
• Kukedlc/NeuralSynthesis-7B-v0.1

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: Kukedlc/NeuralMaths-Experiment-7b
  - model: Kukedlc/NeuralArjuna-7B-DT
  - model: Kukedlc/NeuralSirKrishna-7b
  - model: Kukedlc/NeuralSynthesis-7B-v0.1
merge_method: model_stock
base_model: Kukedlc/NeuralSirKrishna-7b
dtype: bfloat16

Model Inference:

!pip install -qU transformers accelerate bitsandbytes

from transformers import AutoModelForCausalLM, AutoTokenizer, TextStreamer, BitsAndBytesConfig
import torch

bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_use_double_quant=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16
)

MODEL_NAME = 'Kukedlc/NeuralStockFusion-7b'
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
model = AutoModelForCausalLM.from_pretrained(MODEL_NAME, device_map='cuda:0', quantization_config=bnb_config)

inputs = tokenizer(["[INST] What is a large language model, in spanish \n[/INST]\n"], return_tensors="pt").to('cuda')
streamer = TextStreamer(tokenizer)

# Despite returning the usual output, the streamer will also print the generated text to stdout.
_ = model.generate(**inputs, streamer=streamer, max_new_tokens=256, do_sample=True, temperature=0.7, repetition_penalty=1.4, top_p=0.9)