Instructions to use Lernex/Metis-1.6-Praxis with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Lernex/Metis-1.6-Praxis with Transformers:

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

pipe = pipeline("text-generation", model="Lernex/Metis-1.6-Praxis", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Lernex/Metis-1.6-Praxis", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Lernex/Metis-1.6-Praxis with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Lernex/Metis-1.6-Praxis"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Lernex/Metis-1.6-Praxis",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Lernex/Metis-1.6-Praxis

SGLang

How to use Lernex/Metis-1.6-Praxis 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 "Lernex/Metis-1.6-Praxis" \
    --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": "Lernex/Metis-1.6-Praxis",
		"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 "Lernex/Metis-1.6-Praxis" \
        --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": "Lernex/Metis-1.6-Praxis",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Lernex/Metis-1.6-Praxis with Docker Model Runner:
```
docker model run hf.co/Lernex/Metis-1.6-Praxis
```

Metis-1.6-Praxis / NGRAM_MEMORY_REMOVED.json

GiuliannoV

Metis-1.6 Praxis

f6b1be3 18 days ago

Raw History Blame Contribute Delete

2.64 kB

{
 "removed": "ngram_memory.*: hashed tables, projection and gates",
 "source": "/lus/lustre1/vollmerc/metis-1.6/exports/Metis-1.6-Praxis-G80-Full",
 "source_model_safetensors_sha256": "9654d9c86a51215eb15b6fb126c41cfe2cd5a5c0446478a458693735205b83b4",
 "model_safetensors_sha256": "861af5dbaa43923cef0ea4abd71124148ff3078370218fb12d4685fcd1779072",
 "removed_tensors": [
  {
   "name": "ngram_memory.gate_bias",
   "shape": [
    2,
    5,
    4
   ]
  },
  {
   "name": "ngram_memory.gate_vectors",
   "shape": [
    2,
    5,
    4,
    2048
   ]
  },
  {
   "name": "ngram_memory.projection.bias",
   "shape": [
    2048
   ]
  },
  {
   "name": "ngram_memory.projection.weight",
   "shape": [
    2048,
    1024
   ]
  },
  {
   "name": "ngram_memory.tables.o2_h0.embedding.weight",
   "shape": [
    578957,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o2_h1.embedding.weight",
   "shape": [
    579947,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o2_h2.embedding.weight",
   "shape": [
    580939,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o2_h3.embedding.weight",
   "shape": [
    581941,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o2_h4.embedding.weight",
   "shape": [
    582937,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o2_h5.embedding.weight",
   "shape": [
    583937,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o2_h6.embedding.weight",
   "shape": [
    584951,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o2_h7.embedding.weight",
   "shape": [
    585737,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o3_h0.embedding.weight",
   "shape": [
    586139,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o3_h1.embedding.weight",
   "shape": [
    586939,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o3_h2.embedding.weight",
   "shape": [
    587947,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o3_h3.embedding.weight",
   "shape": [
    588937,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o3_h4.embedding.weight",
   "shape": [
    589993,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o3_h5.embedding.weight",
   "shape": [
    590959,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o3_h6.embedding.weight",
   "shape": [
    591937,
    64
   ]
  },
  {
   "name": "ngram_memory.tables.o3_h7.embedding.weight",
   "shape": [
    592793,
    64
   ]
  }
 ],
 "removed_elements": 602180520,
 "kept_tensors": 418,
 "load_with": "ngram_free.without_ngram_memory()",
 "note": "Lernex/Metis-1.6-Praxis: Post-trained overall_sft G80 checkpoint tokens-1022774982313 (8e7512a4), source model.safetensors 9654d9c8"
}

	{
	"removed": "ngram_memory.*: hashed tables, projection and gates",
	"source": "/lus/lustre1/vollmerc/metis-1.6/exports/Metis-1.6-Praxis-G80-Full",
	"source_model_safetensors_sha256": "9654d9c86a51215eb15b6fb126c41cfe2cd5a5c0446478a458693735205b83b4",
	"model_safetensors_sha256": "861af5dbaa43923cef0ea4abd71124148ff3078370218fb12d4685fcd1779072",
	"removed_tensors": [
	{
	"name": "ngram_memory.gate_bias",
	"shape": [
	2,
	5,
	4
	]
	},
	{
	"name": "ngram_memory.gate_vectors",
	"shape": [
	2,
	5,
	4,
	2048
	]
	},
	{
	"name": "ngram_memory.projection.bias",
	"shape": [
	2048
	]
	},
	{
	"name": "ngram_memory.projection.weight",
	"shape": [
	2048,
	1024
	]
	},
	{
	"name": "ngram_memory.tables.o2_h0.embedding.weight",
	"shape": [
	578957,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o2_h1.embedding.weight",
	"shape": [
	579947,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o2_h2.embedding.weight",
	"shape": [
	580939,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o2_h3.embedding.weight",
	"shape": [
	581941,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o2_h4.embedding.weight",
	"shape": [
	582937,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o2_h5.embedding.weight",
	"shape": [
	583937,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o2_h6.embedding.weight",
	"shape": [
	584951,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o2_h7.embedding.weight",
	"shape": [
	585737,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o3_h0.embedding.weight",
	"shape": [
	586139,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o3_h1.embedding.weight",
	"shape": [
	586939,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o3_h2.embedding.weight",
	"shape": [
	587947,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o3_h3.embedding.weight",
	"shape": [
	588937,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o3_h4.embedding.weight",
	"shape": [
	589993,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o3_h5.embedding.weight",
	"shape": [
	590959,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o3_h6.embedding.weight",
	"shape": [
	591937,
	64
	]
	},
	{
	"name": "ngram_memory.tables.o3_h7.embedding.weight",
	"shape": [
	592793,
	64
	]
	}
	],
	"removed_elements": 602180520,
	"kept_tensors": 418,
	"load_with": "ngram_free.without_ngram_memory()",
	"note": "Lernex/Metis-1.6-Praxis: Post-trained overall_sft G80 checkpoint tokens-1022774982313 (8e7512a4), source model.safetensors 9654d9c8"
	}