Instructions to use athirdpath/BigMistral-11b-GLUED with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use athirdpath/BigMistral-11b-GLUED with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="athirdpath/BigMistral-11b-GLUED")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("athirdpath/BigMistral-11b-GLUED")
model = AutoModelForCausalLM.from_pretrained("athirdpath/BigMistral-11b-GLUED")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use athirdpath/BigMistral-11b-GLUED with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "athirdpath/BigMistral-11b-GLUED"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "athirdpath/BigMistral-11b-GLUED",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/athirdpath/BigMistral-11b-GLUED

SGLang

How to use athirdpath/BigMistral-11b-GLUED with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "athirdpath/BigMistral-11b-GLUED" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "athirdpath/BigMistral-11b-GLUED",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "athirdpath/BigMistral-11b-GLUED" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "athirdpath/BigMistral-11b-GLUED",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use athirdpath/BigMistral-11b-GLUED with Docker Model Runner:
```
docker model run hf.co/athirdpath/BigMistral-11b-GLUED
```

athirdpath commited on Dec 4, 2023

Commit

56d9833

1 Parent(s): 1e8448a

Update README.md

Browse files

Files changed (1) hide show

README.md +17 -8

README.md CHANGED Viewed

@@ -1,23 +1,32 @@
 ---
 license: cc-by-nc-4.0
 pipeline_tag: text-generation
 ---
-### Current State
-Well, I've reduced per-step loss by 0.36 (from 1.57 to 1.21) in a third of an epoch. To compare, the (meh) 13b Mistral glue LoRA reduced per-step loss by 0.37 (2.16 to 1.79) over an entire 4 epochs!
-EDIT: First 2 evals both came in at < 1, MUCH better than the 13b attempt. Verdict is that Mistral (or 7Bs in general, I'd guess) can't survive more than one cut, meaningfully. Per-step loss reduced by 0.47 @ 1 epoch.
-EDIT 2: Halfway there. Eval loss < 0.85 for the last 2 evals, promising. Per-step loss down to ~1.07, a reduction of ~33%!
-EDIT 3: 80% done. Curve is greatly flattened, so 3 epochs seems like it was the right call. Eval down to 0.81 and per-step loss down to 0.93. Can't wait to test!
-EDIT 4: Done! Testing time.
 ### Dataset
 The 11b glue consists of:
 - The entirety of HF No Robots.
 - The entirety of TinyPixel/orca-mini
-- Enough of the GPT-4 generated Alpaca dataset (randomly chosen) to make it a roughly even three-way split.

 ---
 license: cc-by-nc-4.0
+language:
+- en
 pipeline_tag: text-generation
 ---
+<p align="center"><font size="7"> <b>Okay, here we fuckin' go.</b> </font></p>
+<p align="center"><font size="5"> <b>Time to fire up the ol' dare_ties pod.</b></font></p>
+<p align="center"><img src="https://iili.io/JzixYiP.png"/>
+<p align="center"><font size="6"><b><a href="https://iili.io/Jzix7WB.png">NSFW - Erotic(?) Writing Example - NSFW</font></a></b></p>
+<p align="center"><font size="3"> <b>(That's not what it's finetuned for, okay? He's a grower.)</b></font></p>
+### Training Log
+1) Well, I've reduced per-step loss by 0.36 (from 1.57 to 1.21) in a third of an epoch. To compare, the (meh) 13b Mistral glue LoRA reduced per-step loss by 0.37 (2.16 to 1.79) over an entire 4 epochs!
+2) First 2 evals both came in at < 1, MUCH better than the 13b attempt. Verdict is that Mistral (or 7Bs in general, I'd guess) can't survive more than one cut, meaningfully. Per-step loss reduced by 0.47 @ 1 epoch.
+3) Halfway there. Eval loss < 0.85 for the last 2 evals, promising. Per-step loss down to ~1.07, a reduction of ~33%!
+4) 80% done. Curve is greatly flattened, so 3 epochs seems like it was the right call. Eval down to 0.81 and per-step loss down to 0.93. Can't wait to test!
+5) Done! Testing time.
 ### Dataset
 The 11b glue consists of:
 - The entirety of HF No Robots.
 - The entirety of TinyPixel/orca-mini
+- Enough of the GPT-4 generated Alpaca dataset (randomly chosen) to make it a roughly even three-way split.
+JSONL file of dataset available as a repo.