Text Generation
Transformers
Safetensors
ncp_olmo3
bfloat16
causal-lm
conceptlm
olmo3
research
custom_code
Instructions to use ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000", trust_remote_code=True)# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000
- SGLang
How to use ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000 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 "ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000" \ --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": "ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000", "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 "ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000" \ --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": "ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000 with Docker Model Runner:
docker model run hf.co/ArchSpace-Collection/NCP_Olmo3_Stage1_Step700000
File size: 4,667 Bytes
a37f751 29b54cc a37f751 29b54cc 1b79d28 29b54cc a37f751 29b54cc 1b79d28 a37f751 29b54cc a37f751 29b54cc 1b79d28 29b54cc 1b79d28 a37f751 29b54cc 1b79d28 29b54cc a37f751 29b54cc 1b79d28 29b54cc 1b79d28 29b54cc 1b79d28 29b54cc a37f751 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 | {
"add_bias_linear": false,
"add_qkv_bias": false,
"apply_residual_connection_post_layernorm": false,
"architectures": [
"NCPOlmo3ForCausalLM"
],
"attention_dropout": 0.0,
"auto_map": {
"AutoConfig": "configuration_ncp_olmo3.NCPOlmo3Config",
"AutoModelForCausalLM": "modeling_ncp_olmo3.NCPOlmo3ForCausalLM"
},
"checkpoint_step": 700000,
"concept_chunk_size": 4,
"concept_shift_feature": true,
"conceptlm_backbone": "olmo3",
"conceptlm_chunk_dualpath_alpha_init": 0.5,
"conceptlm_chunk_merge_method": "meanpooling",
"conceptlm_chunk_size": 4,
"conceptlm_decoder_layers": 16,
"conceptlm_enable_chunk_dualpath_smoothing": false,
"conceptlm_encoder_layers": 16,
"conceptlm_fusion_alpha_init": 1.0,
"conceptlm_fusion_norm_alpha_init": 0.1,
"conceptlm_hlm_attention_mode": "backbone_window",
"conceptlm_hlm_ffn_hidden_size": null,
"conceptlm_hlm_loss_weight": 1.0,
"conceptlm_layer_norm_option": "normed_add",
"conceptlm_shift_feature": true,
"conceptlm_special_layers": 8,
"conceptlm_v21_compile_dd_routes": true,
"conceptlm_v21_compile_residual_flow_routes": true,
"conceptlm_v21_concept_read_encoder_first_n": -1,
"conceptlm_v21_dd_concept_self_dd": true,
"conceptlm_v21_dd_concept_self_dd_every_n_layers": 1,
"conceptlm_v21_dd_concept_self_dd_hidden_size": 0,
"conceptlm_v21_dd_concept_self_dd_use_layernorm": false,
"conceptlm_v21_dd_encoder_self_dd": true,
"conceptlm_v21_dd_encoder_self_dd_every_n_layers": 1,
"conceptlm_v21_dd_encoder_self_dd_hidden_size": 0,
"conceptlm_v21_dd_encoder_self_dd_use_layernorm": false,
"conceptlm_v21_dd_self_dd_mode": "dd",
"conceptlm_v21_dd_two_route_add": true,
"conceptlm_v21_dd_two_route_add_beta_init": 0.05,
"conceptlm_v21_dd_two_route_add_concept_hidden_size": 0,
"conceptlm_v21_dd_two_route_add_concept_route_first_n": -1,
"conceptlm_v21_dd_two_route_add_concept_source": "final",
"conceptlm_v21_dd_two_route_add_concept_use_layernorm": true,
"conceptlm_v21_dd_two_route_add_decoder_hidden_size": 0,
"conceptlm_v21_dd_two_route_add_decoder_use_layernorm": false,
"conceptlm_v21_dd_two_route_add_decoder_use_softmax": true,
"conceptlm_v21_dd_two_route_add_disable_decoder_dd": false,
"conceptlm_v21_dd_two_route_add_enable_final_concept_route": true,
"conceptlm_v21_dd_two_route_add_enable_raw_concept_route": false,
"conceptlm_v21_dd_two_route_add_every_n_layers": 1,
"conceptlm_v21_dd_two_route_add_use_softmax": false,
"conceptlm_v21_decoder_read_encoder_first_n": -1,
"conceptlm_v21_enable_concept_read_encoder": true,
"conceptlm_v21_enable_decoder_read_concept": true,
"conceptlm_v21_enable_decoder_read_encoder": true,
"conceptlm_v21_enable_full_residual_flow": false,
"conceptlm_v21_final_read_concept_gate": true,
"conceptlm_v21_final_read_concept_gate_init_final": 0.5,
"conceptlm_v21_final_read_concept_gate_reg_weight": 0.0,
"conceptlm_v21_final_read_concept_gate_target_final": 0.5,
"conceptlm_v21_residual_flow_beta_init": 0.02,
"conceptlm_v21_residual_flow_route_hidden_size": 0,
"conceptlm_v21_residual_flow_route_use_softmax": true,
"conceptlm_v21_residual_flow_shared_source_norm": true,
"conceptlm_v21_residual_flow_source_use_layernorm": true,
"conceptlm_v22_vq_codebook_size": 128,
"conceptlm_v22_vq_commitment_cost": 0.25,
"conceptlm_v22_vq_detach_hlm_target": true,
"conceptlm_v22_vq_highlevel_lr_mult": 1.0,
"conceptlm_v22_vq_hlm_loss_type": "mse",
"conceptlm_v22_vq_lr_mult": 1.0,
"conceptlm_v22_vq_merge_mode": "raw_logits",
"conceptlm_v22_vq_num_codebooks": 32,
"conceptlm_vq_loss_weight": 1.0,
"custom_model_loader_required": false,
"ffn_hidden_size": 11008,
"fp32_residual_connection": false,
"gated_linear_unit": true,
"hidden_dropout": 0.0,
"hidden_size": 4096,
"kv_channels": 128,
"layernorm_epsilon": 1e-06,
"layernorm_zero_centered_gamma": false,
"max_position_embeddings": 8192,
"max_sequence_length": 8192,
"model_type": "ncp_olmo3",
"normalization": "RMSNorm",
"num_attention_heads": 32,
"num_hidden_layers": 40,
"num_layers": 32,
"num_query_groups": 32,
"params_dtype": "bfloat16",
"position_embedding_type": "rope",
"qk_layernorm": true,
"rotary_base": 500000,
"rotary_interleaved": false,
"rotary_percent": 1.0,
"standalone_hf_backend": true,
"standalone_hf_backend_status": "parity_candidate",
"torch_dtype": "bfloat16",
"transformers_version": "5.8.0",
"untie_embeddings_and_output_weights": true,
"use_cache": true,
"vocab_size": 100278,
"weight_key_format": "native_megatron_state_dict",
"window_attn_skip_freq": 4,
"window_size": [
4096,
0
]
}
|