Text Generation
Safetensors
Transformers
LongCat-Flash-Lite-Sparse
conversational
compressed-tensors
Instructions to use Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import LongcatCausalLM model = LongcatCausalLM.from_pretrained("Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound
- SGLang
How to use Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound 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 "Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound" \ --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": "Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound", "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 "Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound" \ --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": "Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound with Docker Model Runner:
docker model run hf.co/Intel/LongCat-Flash-Lite-Sparse-MXFP4-CT-AutoRound
| { | |
| "architectures": [ | |
| "LongcatCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "vocab_size": 131072, | |
| "hidden_size": 3072, | |
| "ffn_hidden_size": 6144, | |
| "expert_ffn_hidden_size": 1024, | |
| "num_layers": 14, | |
| "num_attention_heads": 32, | |
| "kv_lora_rank": 512, | |
| "q_lora_rank": 1536, | |
| "qk_rope_head_dim": 64, | |
| "v_head_dim": 128, | |
| "qk_nope_head_dim": 128, | |
| "mla_scale_q_lora": true, | |
| "mla_scale_kv_lora": true, | |
| "routed_scaling_factor": 6.0, | |
| "n_routed_experts": 256, | |
| "max_position_embeddings": 983040, | |
| "rms_norm_eps": 1e-05, | |
| "use_cache": true, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "rope_theta": 1000000.0, | |
| "rope_scaling": { | |
| "original_max_position_embeddings": 8192, | |
| "rope_type": "deepseek_yarn", | |
| "factor": 120, | |
| "beta_fast": 32, | |
| "beta_slow": 1, | |
| "mscale": 1, | |
| "mscale_all_dim": 1 | |
| }, | |
| "attention_method": "LSA", | |
| "zero_expert_num": 128, | |
| "zero_expert_type": "identity", | |
| "moe_topk": 12, | |
| "use_mla": 1, | |
| "moe_switch_token_num": 1024, | |
| "moe_impl": "mix", | |
| "oe_vocab_size_ratio": 78, | |
| "oe_neighbor_num": 4, | |
| "oe_split_num": 4, | |
| "mtp_num_layers": 3, | |
| "mtp_replicate_modules": true, | |
| "mtp_disable_over_tokenizer": false, | |
| "indexer_rope_interleave": true, | |
| "index_n_heads": 16, | |
| "index_head_dim": 128, | |
| "index_topk": 2048, | |
| "index_k_norm_type": "rms", | |
| "cli_factor": 2, | |
| "dsa_mtp_cli": true, | |
| "index_local_tokens": 1024, | |
| "index_init_tokens": 16, | |
| "quantization_config": { | |
| "config_groups": { | |
| "group_0": { | |
| "targets": [ | |
| "Linear" | |
| ], | |
| "weights": { | |
| "num_bits": 4, | |
| "type": "float", | |
| "symmetric": true, | |
| "group_size": 32, | |
| "strategy": "group", | |
| "block_structure": null, | |
| "dynamic": false, | |
| "actorder": null, | |
| "scale_dtype": "torch.uint8", | |
| "zp_dtype": null, | |
| "observer": "memoryless_minmax", | |
| "observer_kwargs": {} | |
| }, | |
| "input_activations": { | |
| "num_bits": 4, | |
| "type": "float", | |
| "symmetric": true, | |
| "group_size": 32, | |
| "strategy": "group", | |
| "block_structure": null, | |
| "dynamic": true, | |
| "actorder": null, | |
| "scale_dtype": "torch.uint8", | |
| "zp_dtype": null, | |
| "observer": null, | |
| "observer_kwargs": {} | |
| }, | |
| "output_activations": null, | |
| "format": null | |
| } | |
| }, | |
| "quant_method": "compressed-tensors", | |
| "kv_cache_scheme": null, | |
| "format": "mxfp4-pack-quantized", | |
| "quantization_status": "compressed", | |
| "global_compression_ratio": null, | |
| "ignore": [ | |
| "model.layers.2.self_attn.1.kv_a_proj_with_mqa", | |
| "model.mtp.layers.0.self_attn.kv_a_proj_with_mqa", | |
| "model.oe_embed_proj3", | |
| "model.oe_embed_tokens10", | |
| "model.layers.0.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.10.self_attn.0.kv_a_proj_with_mqa", | |
| "model.oe_embed_proj8", | |
| "model.oe_embed_tokens11", | |
| "model.layers.1.self_attn.0.indexer.wk", | |
| "model.layers.9.mlp.router.classifier", | |
| "model.oe_embed_proj0", | |
| "model.oe_embed_proj4", | |
| "model.oe_embed_tokens9", | |
| "model.layers.3.self_attn.0.indexer.weights_proj", | |
| "model.layers.8.mlp.router.classifier", | |
| "model.layers.9.self_attn.1.kv_a_proj_with_mqa", | |
| "model.oe_embed_tokens8", | |
| "model.layers.10.self_attn.0.indexer.weights_proj", | |
| "model.layers.7.mlp.router.classifier", | |
| "model.layers.8.self_attn.1.kv_a_proj_with_mqa", | |
| "model.oe_embed_tokens7", | |
| "model.layers.2.self_attn.0.indexer.weights_proj", | |
| "model.layers.6.mlp.router.classifier", | |
| "model.layers.7.self_attn.1.kv_a_proj_with_mqa", | |
| "model.oe_embed_tokens6", | |
| "model.layers.13.self_attn.0.indexer.weights_proj", | |
| "model.layers.5.mlp.router.classifier", | |
| "model.layers.6.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.9.self_attn.0.indexer.weights_proj", | |
| "model.oe_embed_tokens5", | |
| "model.layers.12.self_attn.0.indexer.weights_proj", | |
| "model.layers.4.mlp.router.classifier", | |
| "model.layers.5.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.8.self_attn.0.indexer.weights_proj", | |
| "model.oe_embed_tokens4", | |
| "model.layers.11.self_attn.0.indexer.weights_proj", | |
| "model.layers.3.mlp.router.classifier", | |
| "model.layers.4.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.7.self_attn.0.indexer.weights_proj", | |
| "model.oe_embed_tokens3", | |
| "model.layers.1.self_attn.0.indexer.weights_proj", | |
| "model.layers.10.mlp.router.classifier", | |
| "model.layers.3.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.6.self_attn.0.indexer.weights_proj", | |
| "model.oe_embed_tokens2", | |
| "model.layers.0.self_attn.0.indexer.weights_proj", | |
| "model.layers.10.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.2.mlp.router.classifier", | |
| "model.layers.5.self_attn.0.indexer.weights_proj", | |
| "model.oe_embed_tokens1", | |
| "model.layers.13.mlp.router.classifier", | |
| "model.layers.4.self_attn.0.indexer.weights_proj", | |
| "model.layers.9.self_attn.0.kv_a_proj_with_mqa", | |
| "model.mtp.layers.0.self_attn.indexer.weights_proj", | |
| "model.oe_embed_tokens0", | |
| "model.embed_tokens", | |
| "model.layers.1.mlp.router.classifier", | |
| "model.layers.12.self_attn.0.indexer.wq_b", | |
| "model.layers.13.self_attn.0.indexer.wk", | |
| "model.layers.6.self_attn.0.indexer.wq_b", | |
| "model.layers.6.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.11.mlp.router.classifier", | |
| "model.layers.13.self_attn.0.indexer.wq_b", | |
| "model.layers.2.self_attn.0.indexer.wk", | |
| "model.layers.7.self_attn.0.indexer.wq_b", | |
| "model.layers.7.self_attn.0.kv_a_proj_with_mqa", | |
| "model.mtp.embed_tokens", | |
| "lm_head", | |
| "model.layers.10.self_attn.0.indexer.wk", | |
| "model.layers.12.mlp.router.classifier", | |
| "model.layers.2.self_attn.0.indexer.wq_b", | |
| "model.layers.8.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.0.mlps.1.gate_proj", | |
| "model.layers.0.self_attn.0.o_proj", | |
| "model.layers.1.self_attn.0.q_a_proj", | |
| "model.layers.1.self_attn.1.q_b_proj", | |
| "model.layers.10.self_attn.0.kv_b_proj", | |
| "model.layers.11.self_attn.1.kv_b_proj", | |
| "model.layers.12.self_attn.1.q_a_proj", | |
| "model.layers.13.mlps.1.down_proj", | |
| "model.layers.13.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.2.mlps.1.up_proj", | |
| "model.layers.2.self_attn.1.o_proj", | |
| "model.layers.3.self_attn.0.indexer.wk", | |
| "model.layers.3.self_attn.1.o_proj", | |
| "model.layers.4.mlps.1.down_proj", | |
| "model.layers.4.mlps.1.up_proj", | |
| "model.layers.5.self_attn.1.kv_b_proj", | |
| "model.layers.6.mlps.0.up_proj", | |
| "model.layers.8.self_attn.1.q_b_proj", | |
| "model.layers.9.mlps.0.down_proj", | |
| "model.layers.9.self_attn.0.q_a_proj", | |
| "model.mtp.layers.0.transformer_layer.mlp.down_proj", | |
| "model.oe_embed_proj10", | |
| "model.layers.0.mlps.0.down_proj", | |
| "model.layers.1.mlps.1.gate_proj", | |
| "model.layers.1.self_attn.0.o_proj", | |
| "model.layers.10.mlps.1.gate_proj", | |
| "model.layers.10.self_attn.1.q_a_proj", | |
| "model.layers.11.self_attn.0.q_a_proj", | |
| "model.layers.11.self_attn.1.q_b_proj", | |
| "model.layers.12.self_attn.1.kv_b_proj", | |
| "model.layers.13.self_attn.1.q_a_proj", | |
| "model.layers.2.mlps.0.gate_proj", | |
| "model.layers.2.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.3.self_attn.0.kv_b_proj", | |
| "model.layers.4.self_attn.0.indexer.wk", | |
| "model.layers.4.self_attn.1.o_proj", | |
| "model.layers.5.mlps.0.gate_proj", | |
| "model.layers.5.mlps.1.up_proj", | |
| "model.layers.6.self_attn.1.kv_b_proj", | |
| "model.layers.7.mlps.0.up_proj", | |
| "model.layers.9.self_attn.1.q_b_proj", | |
| "model.mtp.layers.0.self_attn.o_proj", | |
| "model.mtp.layers.0.transformer_layer.mlp.gate_proj", | |
| "model.oe_embed_proj5", | |
| "model.layers.0.mlps.0.gate_proj", | |
| "model.layers.0.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.1.mlps.0.down_proj", | |
| "model.layers.10.self_attn.0.o_proj", | |
| "model.layers.10.self_attn.0.q_b_proj", | |
| "model.layers.11.mlps.1.gate_proj", | |
| "model.layers.11.self_attn.0.o_proj", | |
| "model.layers.12.self_attn.0.q_a_proj", | |
| "model.layers.12.self_attn.1.q_b_proj", | |
| "model.layers.13.self_attn.1.kv_b_proj", | |
| "model.layers.2.mlps.1.down_proj", | |
| "model.layers.2.self_attn.1.q_a_proj", | |
| "model.layers.3.mlps.1.gate_proj", | |
| "model.layers.3.self_attn.1.q_a_proj", | |
| "model.layers.4.self_attn.0.kv_b_proj", | |
| "model.layers.5.mlps.1.down_proj", | |
| "model.layers.5.self_attn.0.indexer.wk", | |
| "model.layers.5.self_attn.1.o_proj", | |
| "model.layers.6.mlps.1.up_proj", | |
| "model.layers.7.self_attn.1.kv_b_proj", | |
| "model.layers.8.mlps.0.up_proj", | |
| "model.oe_embed_proj1", | |
| "model.layers.0.mlps.1.down_proj", | |
| "model.layers.1.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.11.mlps.0.down_proj", | |
| "model.layers.12.mlps.1.gate_proj", | |
| "model.layers.12.self_attn.0.o_proj", | |
| "model.layers.13.self_attn.0.q_a_proj", | |
| "model.layers.13.self_attn.1.q_b_proj", | |
| "model.layers.2.self_attn.1.kv_b_proj", | |
| "model.layers.3.self_attn.0.o_proj", | |
| "model.layers.3.self_attn.0.q_b_proj", | |
| "model.layers.4.mlps.1.gate_proj", | |
| "model.layers.4.self_attn.1.q_a_proj", | |
| "model.layers.5.self_attn.0.kv_b_proj", | |
| "model.layers.6.mlps.0.gate_proj", | |
| "model.layers.6.self_attn.0.indexer.wk", | |
| "model.layers.6.self_attn.1.o_proj", | |
| "model.layers.7.mlps.1.up_proj", | |
| "model.layers.8.self_attn.1.kv_b_proj", | |
| "model.layers.9.mlps.0.up_proj", | |
| "model.mtp.layers.0.eh_proj", | |
| "model.mtp.layers.0.self_attn.q_a_proj", | |
| "model.oe_embed_proj11", | |
| "model.layers.1.mlps.0.gate_proj", | |
| "model.layers.10.mlps.0.down_proj", | |
| "model.layers.10.self_attn.0.q_a_proj", | |
| "model.layers.11.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.12.mlps.0.down_proj", | |
| "model.layers.13.mlps.1.gate_proj", | |
| "model.layers.13.self_attn.0.o_proj", | |
| "model.layers.2.self_attn.0.q_a_proj", | |
| "model.layers.2.self_attn.1.q_b_proj", | |
| "model.layers.4.self_attn.0.o_proj", | |
| "model.layers.4.self_attn.0.q_b_proj", | |
| "model.layers.5.mlps.1.gate_proj", | |
| "model.layers.5.self_attn.1.q_a_proj", | |
| "model.layers.6.mlps.1.down_proj", | |
| "model.layers.6.self_attn.0.kv_b_proj", | |
| "model.layers.7.self_attn.0.indexer.wk", | |
| "model.layers.7.self_attn.1.o_proj", | |
| "model.layers.8.mlps.1.up_proj", | |
| "model.layers.9.self_attn.1.kv_b_proj", | |
| "model.mtp.layers.0.self_attn.kv_b_proj", | |
| "model.mtp.layers.0.transformer_layer.mlp.up_proj", | |
| "model.oe_embed_proj6", | |
| "model.layers.0.mlps.0.up_proj", | |
| "model.layers.0.self_attn.0.kv_b_proj", | |
| "model.layers.0.self_attn.1.q_a_proj", | |
| "model.layers.1.mlps.1.down_proj", | |
| "model.layers.1.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.10.self_attn.1.q_b_proj", | |
| "model.layers.11.self_attn.0.indexer.wk", | |
| "model.layers.13.mlps.0.down_proj", | |
| "model.layers.2.mlps.1.gate_proj", | |
| "model.layers.2.self_attn.0.o_proj", | |
| "model.layers.3.mlps.0.down_proj", | |
| "model.layers.3.self_attn.0.q_a_proj", | |
| "model.layers.4.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.5.self_attn.0.o_proj", | |
| "model.layers.5.self_attn.0.q_b_proj", | |
| "model.layers.6.mlps.1.gate_proj", | |
| "model.layers.6.self_attn.1.q_a_proj", | |
| "model.layers.7.mlps.0.gate_proj", | |
| "model.layers.7.self_attn.0.kv_b_proj", | |
| "model.layers.8.self_attn.1.o_proj", | |
| "model.layers.9.mlps.1.up_proj", | |
| "model.layers.0.mlps.1.up_proj", | |
| "model.layers.0.self_attn.1.o_proj", | |
| "model.layers.1.mlps.0.up_proj", | |
| "model.layers.1.self_attn.0.kv_b_proj", | |
| "model.layers.1.self_attn.1.q_a_proj", | |
| "model.layers.10.mlps.0.gate_proj", | |
| "model.layers.11.mlps.0.gate_proj", | |
| "model.layers.11.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.12.self_attn.0.indexer.wk", | |
| "model.layers.2.mlps.0.down_proj", | |
| "model.layers.3.self_attn.1.q_b_proj", | |
| "model.layers.4.mlps.0.down_proj", | |
| "model.layers.4.self_attn.0.q_a_proj", | |
| "model.layers.5.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.6.self_attn.0.o_proj", | |
| "model.layers.6.self_attn.0.q_b_proj", | |
| "model.layers.7.mlps.1.down_proj", | |
| "model.layers.7.mlps.1.gate_proj", | |
| "model.layers.7.self_attn.1.q_a_proj", | |
| "model.layers.8.self_attn.0.kv_b_proj", | |
| "model.layers.9.self_attn.1.o_proj", | |
| "model.layers.1.mlps.1.up_proj", | |
| "model.layers.1.self_attn.1.o_proj", | |
| "model.layers.10.mlps.0.up_proj", | |
| "model.layers.10.mlps.1.down_proj", | |
| "model.layers.10.self_attn.0.indexer.wq_b", | |
| "model.layers.11.mlps.0.up_proj", | |
| "model.layers.11.mlps.1.down_proj", | |
| "model.layers.12.self_attn.0.q_b_proj", | |
| "model.layers.13.self_attn.0.kv_b_proj", | |
| "model.layers.3.self_attn.0.indexer.wq_b", | |
| "model.layers.3.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.3.self_attn.1.kv_b_proj", | |
| "model.layers.4.self_attn.1.q_b_proj", | |
| "model.layers.5.mlps.0.down_proj", | |
| "model.layers.5.self_attn.0.q_a_proj", | |
| "model.layers.7.self_attn.0.o_proj", | |
| "model.layers.7.self_attn.0.q_b_proj", | |
| "model.layers.8.mlps.0.gate_proj", | |
| "model.layers.8.mlps.1.gate_proj", | |
| "model.layers.8.self_attn.1.q_a_proj", | |
| "model.layers.9.self_attn.0.indexer.wq_b", | |
| "model.mtp.layers.0.self_attn.indexer.wk", | |
| "model.mtp.layers.0.self_attn.indexer.wq_b", | |
| "model.mtp.layers.0.self_attn.q_b_proj", | |
| "model.layers.0.mlp.router.classifier", | |
| "model.layers.0.self_attn.0.indexer.wk", | |
| "model.layers.0.self_attn.0.indexer.wq_b", | |
| "model.layers.0.self_attn.0.q_b_proj", | |
| "model.layers.1.self_attn.0.indexer.wq_b", | |
| "model.layers.11.mlps.1.up_proj", | |
| "model.layers.11.self_attn.0.indexer.wq_b", | |
| "model.layers.11.self_attn.1.o_proj", | |
| "model.layers.12.mlps.0.gate_proj", | |
| "model.layers.12.mlps.0.up_proj", | |
| "model.layers.12.self_attn.0.kv_a_proj_with_mqa", | |
| "model.layers.13.self_attn.0.q_b_proj", | |
| "model.layers.2.self_attn.0.kv_b_proj", | |
| "model.layers.3.mlps.0.gate_proj", | |
| "model.layers.3.mlps.0.up_proj", | |
| "model.layers.4.self_attn.0.indexer.wq_b", | |
| "model.layers.4.self_attn.1.kv_b_proj", | |
| "model.layers.5.self_attn.0.indexer.wq_b", | |
| "model.layers.5.self_attn.1.q_b_proj", | |
| "model.layers.6.mlps.0.down_proj", | |
| "model.layers.6.self_attn.0.q_a_proj", | |
| "model.layers.8.mlps.1.down_proj", | |
| "model.layers.8.self_attn.0.indexer.wq_b", | |
| "model.layers.8.self_attn.0.o_proj", | |
| "model.layers.8.self_attn.0.q_b_proj", | |
| "model.layers.9.mlps.1.gate_proj", | |
| "model.layers.9.self_attn.1.q_a_proj", | |
| "model.oe_embed_proj9", | |
| "model.layers.0.self_attn.1.kv_b_proj", | |
| "model.layers.1.self_attn.0.q_b_proj", | |
| "model.layers.10.mlps.1.up_proj", | |
| "model.layers.11.self_attn.0.kv_b_proj", | |
| "model.layers.12.mlps.1.down_proj", | |
| "model.layers.12.mlps.1.up_proj", | |
| "model.layers.12.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.12.self_attn.1.o_proj", | |
| "model.layers.13.mlps.0.up_proj", | |
| "model.layers.2.self_attn.0.q_b_proj", | |
| "model.layers.3.mlps.1.down_proj", | |
| "model.layers.4.mlps.0.up_proj", | |
| "model.layers.6.self_attn.1.q_b_proj", | |
| "model.layers.7.mlps.0.down_proj", | |
| "model.layers.7.self_attn.0.q_a_proj", | |
| "model.layers.8.self_attn.0.indexer.wk", | |
| "model.layers.9.mlps.0.gate_proj", | |
| "model.layers.9.self_attn.0.kv_b_proj", | |
| "model.layers.9.self_attn.0.o_proj", | |
| "model.layers.9.self_attn.0.q_b_proj", | |
| "model.oe_embed_proj2", | |
| "model.layers.0.self_attn.0.q_a_proj", | |
| "model.layers.0.self_attn.1.q_b_proj", | |
| "model.layers.1.self_attn.1.kv_b_proj", | |
| "model.layers.10.self_attn.1.kv_b_proj", | |
| "model.layers.10.self_attn.1.o_proj", | |
| "model.layers.11.self_attn.0.q_b_proj", | |
| "model.layers.11.self_attn.1.q_a_proj", | |
| "model.layers.12.self_attn.0.kv_b_proj", | |
| "model.layers.13.mlps.0.gate_proj", | |
| "model.layers.13.mlps.1.up_proj", | |
| "model.layers.13.self_attn.1.kv_a_proj_with_mqa", | |
| "model.layers.13.self_attn.1.o_proj", | |
| "model.layers.2.mlps.0.up_proj", | |
| "model.layers.3.mlps.1.up_proj", | |
| "model.layers.4.mlps.0.gate_proj", | |
| "model.layers.5.mlps.0.up_proj", | |
| "model.layers.7.self_attn.1.q_b_proj", | |
| "model.layers.8.mlps.0.down_proj", | |
| "model.layers.8.self_attn.0.q_a_proj", | |
| "model.layers.9.mlps.1.down_proj", | |
| "model.layers.9.self_attn.0.indexer.wk", | |
| "model.oe_embed_proj7" | |
| ], | |
| "provider": "auto-round" | |
| } | |
| } |