Instructions to use OpenOneRec/KSA-V2-1.9B-base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OpenOneRec/KSA-V2-1.9B-base with Transformers:

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

pipe = pipeline("text-generation", model="OpenOneRec/KSA-V2-1.9B-base", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OpenOneRec/KSA-V2-1.9B-base", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("OpenOneRec/KSA-V2-1.9B-base", trust_remote_code=True, device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OpenOneRec/KSA-V2-1.9B-base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OpenOneRec/KSA-V2-1.9B-base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OpenOneRec/KSA-V2-1.9B-base",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/OpenOneRec/KSA-V2-1.9B-base

SGLang

How to use OpenOneRec/KSA-V2-1.9B-base 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 "OpenOneRec/KSA-V2-1.9B-base" \
    --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": "OpenOneRec/KSA-V2-1.9B-base",
		"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 "OpenOneRec/KSA-V2-1.9B-base" \
        --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": "OpenOneRec/KSA-V2-1.9B-base",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use OpenOneRec/KSA-V2-1.9B-base with Docker Model Runner:
```
docker model run hf.co/OpenOneRec/KSA-V2-1.9B-base
```

KSA-V2-1.9B-base

File size: 2,467 Bytes

03c3dff

{
  "architectures": [
    "Qwen3ForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoModel": "modeling_qwen3.Qwen3Model",
    "AutoModelForCausalLM": "modeling_qwen3.Qwen3ForCausalLM"
  },
  "bos_token_id": 151643,
  "eos_token_id": 151643,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 6144,
  "max_position_embeddings": 131072,
  "max_window_layers": 24,
  "mix_coeff": 0.0,
  "model_type": "qwen3",
  "num_attention_heads": 16,
  "num_hidden_layers": 24,
  "num_key_value_heads": 16,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000,
  "sksa_apply_sparse_attention": true,
  "sksa_enabled": true,
  "sksa_indexer_block_size": 4,
  "sksa_indexer_chunk_attention_include_summary": false,
  "sksa_indexer_chunk_attention_qk_source": "independent",
  "sksa_indexer_chunk_attention_v_source": "independent",
  "sksa_indexer_force_sink": true,
  "sksa_indexer_head_dim": 128,
  "sksa_indexer_heads": 8,
  "sksa_indexer_qk_norm": false,
  "sksa_indexer_query_block_size": 512,
  "sksa_indexer_query_rank": 512,
  "sksa_indexer_rope_dim": 64,
  "sksa_indexer_scorer_backend": "fused",
  "sksa_indexer_summary_key_source": "chunk_attention",
  "sksa_indexer_teacher_pooling": "max",
  "sksa_indexer_token_budget": 4096,
  "sksa_local_window_tokens": 1024,
  "sksa_routing_mode": "summary_indexer",
  "sksa_sparse_full_layers": [
    3,
    7,
    11,
    15,
    19,
    23
  ],
  "sliding_window": null,
  "summary_chunk_position_ids_type": "origin",
  "summary_chunk_size": 8,
  "summary_independent_attention_layernorm": false,
  "summary_independent_parameters": true,
  "summary_layer_freq": "([1]*4)*6",
  "summary_meta_pool_block_size": 512,
  "summary_meta_pool_group_size": "([128]*1+[128]*1+[128]*1+[0]*1)*6",
  "summary_meta_pool_mass_correction": "log_group_size",
  "summary_meta_pool_mode": "dimension_wise",
  "summary_meta_pool_position": "group_last",
  "summary_partial_chunk_policy": "defer",
  "summary_sliding_chunk_num": "([128]*3+[16384]*1)*6",
  "summary_token_begin": 151936,
  "summary_token_num": 1,
  "summary_token_position_ids_type": "last_chunk_slice_right",
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "5.0.0",
  "truncate_predict_nums": 151936,
  "use_cache": true,
  "use_sliding_window": false,
  "use_summary_attention": true,
  "vocab_size": 151937
}