Update README.md

README.md

@@ -1,45 +1,83 @@
 ---
-base_model: GeneZC/MiniChat-2-3B
+base_model: NousResearch/Nous-Hermes-2-SOLAR-10.7B
 inference: True
 model_type: Llama
+tags:
+- nm-vllm
+- sparse
 ---
-# Nous-Hermes-2-SOLAR-10.7B
-This repo contains pruned model files for [Nous-Hermes-2-SOLAR-10.7B](https://huggingface.co/NousResearch/Nous-Hermes-2-SOLAR-10.7B).
+## Nous-Hermes-2-SOLAR-10.7B-pruned2.4
+This repo contains model files for [Nous Hermes 2 - Yi-34B](https://huggingface.co/NousResearch/Nous-Hermes-2-Yi-34B) optimized for [NM-vLLM](https://github.com/neuralmagic/nm-vllm), a high-throughput serving engine for compressed LLMs.
 
 This model was pruned with [SparseGPT](https://arxiv.org/abs/2301.00774), using [SparseML](https://github.com/neuralmagic/sparseml).
+
+## Inference
+Install [NM-vLLM](https://github.com/neuralmagic/nm-vllm) for fast inference and low memory-usage: 
+```bash
+pip install nm-vllm[sparse]
+```
+Run in a Python pipeline for local inference:
 ```python
-import torch
-from transformers import AutoTokenizer, AutoModelForCausalLM
+from vllm import LLM, SamplingParams
+
+model = LLM("Nous-Hermes-2-SOLAR-10.7B-pruned2.4", sparsity="sparse_w16a16")
 prompt = "How to make banana bread?"
-formatted_prompt =  f"### User:\n{prompt}\n\n### Assistant:\n"
-model_id = "nm-testing/Nous-Hermes-2-SOLAR-10.7B-pruned50-24"
-model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", torch_dtype=torch.float16)
-tokenizer = AutoTokenizer.from_pretrained(model_id)
-inputs = tokenizer(formatted_prompt, return_tensors="pt")
-outputs = model.generate(**inputs, max_new_tokens=200)
-print(tokenizer.batch_decode(outputs)[0])
+formatted_prompt =  f"<|im_start|>User:{prompt}\n<|im_start|>assistant:\n"
+
+sampling_params = SamplingParams(max_tokens=100)
+outputs = model.generate(formatted_prompt, sampling_params=sampling_params)
+print(outputs[0].outputs[0].text)
+"""
+To make banana bread, you will need to follow these steps:
+
+1. Gather the ingredients needed for the recipe: flour, sugar, eggs, baking powder, and banana bread.
+2. Preparing the bread by kneading the flour, butter, and sugar in a bowl until it becomes soft.
+3. Then, knead the bread with the flour eggs and sugar then bake them in a microwave oven.
+4. Once done baking a
 """
-<s> ### User:
-How to make banana bread?
-
-### Assistant:
-To make banana bread, you will need the following ingredients:
-
-- 1 cup of flour
-- 1 cup of sugar
-- 3 cups of bananas
-- 1 cup of milk
-- 1 cup of eggs
-- 1 cup of baking powder
-
-Instructions:
-
-1. Mix the flour and sugar together.
-2. Add the bananas to the mixture.
-3. Add the milk and eggs to the mixture.
-4. Add the baking powder to the mixture.
-5. Mix all the ingredients together.
-6. Pour the mixture into a baking pan.
-7. Bake the mixture for 30 minutes at 350 degrees.<|im_end|>
+```
+
+## Prompt template
+
+```
+### User:\n{prompt}\n\n### Assistant:\n"
+```
+
+## Sparsification
+For details on how this model was sparsified, see the `recipe.yaml` in this repo and follow the instructions below.
+
+Install [SparseML](https://github.com/neuralmagic/sparseml):
+```bash
+git clone https://github.com/neuralmagic/sparseml
+pip install -e "sparseml[transformers]"
+```
+
+Replace the recipe as you like and run this one-shot compression script to apply SparseGPT:
+```python
+import sparseml.transformers
+
+original_model_name = "NousResearch/Nous-Hermes-2-SOLAR-10.7B"
+calibration_dataset = "open_platypus"
+output_directory = "output/"
+
+recipe = """
+test_stage:
+  obcq_modifiers:
+    SparseGPTModifier:
+      sparsity: 0.5
+      sequential_update: true
+      targets: ['re:model.layers.\d*$']
 """
-```
+
+# Apply SparseGPT to the model
+sparseml.transformers.oneshot(
+    model=original_model_name,
+    dataset=calibration_dataset,
+    recipe=recipe,
+    output_dir=output_directory,
+)
+```
+
+## Slack
+
+For further support, and discussions on these models and AI in general, join [Neural Magic's Slack Community](https://join.slack.com/t/discuss-neuralmagic/shared_invite/zt-q1a1cnvo-YBoICSIw3L1dmQpjBeDurQ)