Instructions to use nitrai-research/Moderato-V1-Pro with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nitrai-research/Moderato-V1-Pro with Transformers:

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

pipe = pipeline("text-generation", model="nitrai-research/Moderato-V1-Pro")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("nitrai-research/Moderato-V1-Pro", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nitrai-research/Moderato-V1-Pro with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "nitrai-research/Moderato-V1-Pro"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nitrai-research/Moderato-V1-Pro",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/nitrai-research/Moderato-V1-Pro

SGLang

How to use nitrai-research/Moderato-V1-Pro 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 "nitrai-research/Moderato-V1-Pro" \
    --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": "nitrai-research/Moderato-V1-Pro",
		"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 "nitrai-research/Moderato-V1-Pro" \
        --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": "nitrai-research/Moderato-V1-Pro",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use nitrai-research/Moderato-V1-Pro with Docker Model Runner:
```
docker model run hf.co/nitrai-research/Moderato-V1-Pro
```

Moderato-V1-Pro / config.json

NitrAI

Sync parameter naming to exact physical 113.3B count (config.json)

9dcf818 verified 5 days ago

Raw

History Blame Contribute Delete

5.32 kB

	{
	"architectures": [
	"ModeratoRRRMoeForCausalLM"
	],
	"image_token_id": 248056,
	"language_model_only": false,
	"model_type": "moderato_moe",
	"text_config": {
	"attention_bias": false,
	"attention_dropout": 0.0,
	"attn_output_gate": true,
	"bos_token_id": 248044,
	"dtype": "bfloat16",
	"eos_token_id": 248044,
	"full_attention_interval": 4,
	"head_dim": 256,
	"hidden_act": "silu",
	"hidden_size": 5120,
	"initializer_range": 0.02,
	"intermediate_size": 17408,
	"layer_types": [
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"full_attention"
	],
	"linear_conv_kernel_dim": 4,
	"linear_key_head_dim": 128,
	"linear_num_key_heads": 16,
	"linear_num_value_heads": 48,
	"linear_value_head_dim": 128,
	"mamba_ssm_dtype": "float32",
	"max_position_embeddings": 262144,
	"model_type": "qwen3_5_text",
	"mtp_num_hidden_layers": 1,
	"mtp_use_dedicated_embeddings": false,
	"num_attention_heads": 24,
	"num_hidden_layers": 64,
	"num_key_value_heads": 4,
	"output_gate_type": "swish",
	"pad_token_id": null,
	"partial_rotary_factor": 0.25,
	"rms_norm_eps": 1e-06,
	"rope_parameters": {
	"mrope_interleaved": true,
	"mrope_section": [
	11,
	11,
	10
	],
	"partial_rotary_factor": 0.25,
	"rope_theta": 10000000,
	"rope_type": "default"
	},
	"tie_word_embeddings": false,
	"use_cache": true,
	"vocab_size": 248320
	},
	"tie_word_embeddings": false,
	"transformers_version": "5.8.0.dev0",
	"video_token_id": 248057,
	"vision_config": {
	"deepstack_visual_indexes": [],
	"depth": 27,
	"hidden_act": "gelu_pytorch_tanh",
	"hidden_size": 1152,
	"in_channels": 3,
	"initializer_range": 0.02,
	"intermediate_size": 4304,
	"model_type": "qwen3_5",
	"num_heads": 16,
	"num_position_embeddings": 2304,
	"out_hidden_size": 5120,
	"patch_size": 16,
	"spatial_merge_size": 2,
	"temporal_patch_size": 2
	},
	"vision_end_token_id": 248054,
	"vision_start_token_id": 248053,
	"num_experts": 6,
	"num_experts_per_tok": 2,
	"total_parameters": "113.3B",
	"active_parameters": "32.7B",
	"rrr_technology": {
	"enabled": true,
	"version": "1.0",
	"level1_static_moe_gate": "Softmax(TopK(Wg * x + eps, k=2))",
	"level2_divergence_probe": "Checkpointed Divergence Probe p_theta(h_t, g)",
	"divergence_interval_tokens": 64,
	"divergence_threshold": 0.3,
	"confidence_threshold": 0.5,
	"state_transitions": {
	"continue": "delta < 0.3 (Fast path)",
	"redirect": "delta >= 0.3, c >= 0.5 (Hot-swap specialized expert without context loss)",
	"escalate": "c < 0.5 (Escape to meta-orchestrator)"
	},
	"experts": [
	{
	"id": 0,
	"name": "anti_bloat",
	"specialization": "Clean concise code without boilerplate"
	},
	{
	"id": 1,
	"name": "clean_diffs",
	"specialization": "Git unified diff patches and surgical edits"
	},
	{
	"id": 2,
	"name": "deep_math_cot",
	"specialization": "Complex mathematical reasoning and CoT"
	},
	{
	"id": 3,
	"name": "systems_rust",
	"specialization": "Low-level systems, memory safety and Rust"
	},
	{
	"id": 4,
	"name": "modern_apis",
	"specialization": "Modern SWE APIs, frameworks and async"
	},
	{
	"id": 5,
	"name": "agentic_fable",
	"specialization": "Autonomous agent planning and reflection"
	}
	]
	}
	}