Instructions to use Retreatcost/Evertide-RX-12B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Retreatcost/Evertide-RX-12B with Transformers:

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

pipe = pipeline("text-generation", model="Retreatcost/Evertide-RX-12B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Retreatcost/Evertide-RX-12B")
model = AutoModelForCausalLM.from_pretrained("Retreatcost/Evertide-RX-12B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Retreatcost/Evertide-RX-12B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Retreatcost/Evertide-RX-12B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Retreatcost/Evertide-RX-12B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Retreatcost/Evertide-RX-12B

SGLang

How to use Retreatcost/Evertide-RX-12B 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 "Retreatcost/Evertide-RX-12B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Retreatcost/Evertide-RX-12B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "Retreatcost/Evertide-RX-12B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Retreatcost/Evertide-RX-12B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Retreatcost/Evertide-RX-12B with Docker Model Runner:
```
docker model run hf.co/Retreatcost/Evertide-RX-12B
```

Retreatcost commited on 3 days ago

Commit

dbe8e32

verified ·

1 Parent(s): 59322a6

Update README.md

Browse files

Files changed (1) hide show

README.md +17 -4

README.md CHANGED Viewed

@@ -24,7 +24,7 @@ license: apache-2.0
 ![evertide_rx](https://cdn-uploads.huggingface.co/production/uploads/6671dd5203d6e8087aaf7ce5/zTuxJU9fwrkFbCvkGW1qe.jpeg)
-A generalist model, with some reasoning capabilities and some multi-lang support.
 Supported languages:
 - French
@@ -72,10 +72,17 @@ I trained 2 variants of the model:
   - with unrolled turns (each turn in separate sample)
   - with regular turns (all turns in single sample)
-Unrolled turns teach local attention much better and train faster, but generalize worse for multi-turn (LA, Local attention).
-Regular turns have much better multi-turn generalisation, but they tend to memorize instead of training new capabilities. (GA, Global attention)
-Then I merged these training runs in a pattern 4:1, similar to how Gemma models have layered SWA and GA.
 The following YAML configuration was used to produce this model:
@@ -155,4 +162,10 @@ For this model achieving ERP capabilities wasn't the goal, so I'm happy with cur
 ### RP/ERP model when?
 Soon™.
 </details>

 ![evertide_rx](https://cdn-uploads.huggingface.co/production/uploads/6671dd5203d6e8087aaf7ce5/zTuxJU9fwrkFbCvkGW1qe.jpeg)
+A generalist model, with some reasoning capabilities and multi-lang support.
 Supported languages:
 - French
   - with unrolled turns (each turn in separate sample)
   - with regular turns (all turns in single sample)
+Unrolled turns teach local attention much better and train faster, but generalize worse for multi-turn (Evertide-LA-12B, Local attention).
+Regular turns have much better multi-turn generalisation, but they tend to memorize instead of training new capabilities. (Evertide-GA-12B, Global attention).
+I also trained these with changed RoPE theta - 10K for GA, 10M for LA.
+My reasoning behind this is that during merging I "unrotate" the changes in config, effectively creating a distribution that I haven't trained in.
+LA becomes shrinked to be even more specialized in short context, while GA gets stretched to cover longer context.
+Then I merged these training runs using passthrough in a pattern 4:1, similar to how Gemma 4 models have layered SWA and GA.
+![download](https://cdn-uploads.huggingface.co/production/uploads/6671dd5203d6e8087aaf7ce5/9M_XguM0q7Pv66X8Vy8t9.jpeg)
 The following YAML configuration was used to produce this model:
 ### RP/ERP model when?
 Soon™.
+### Did you train in RL?
+No, not yet, but that's one of future plans.
+### Is the reasoning performative?
+It's hard to tell exactly, it definitely has some elements of it, but it also was trainded with some specific constraints, that force causality between thinking blocks and answer. So I would say that it's at least a hybrid. Any further improvements require RL training.
 </details>