Index

Symbols

--adapter-file ilab-model-convert command line option ilab-model-test command line option --api-key ilab-data-generate command line option ilab-model-chat command line option ilab-model-test command line option --backend ilab-model-evaluate command line option ilab-model-serve command line option --base-branch ilab-model-evaluate command line option --base-model ilab-model-evaluate command line option --batch-size ilab-model-evaluate command line option --benchmark ilab-model-evaluate command line option --branch ilab-model-evaluate command line option --chat-template ilab-model-serve command line option --chunk-word-count ilab-data-generate command line option --ckpt-output-dir ilab-model-train command line option --config ilab command line option --context ilab-model-chat command line option --data-dir ilab-model-test command line option --data-output-dir ilab-model-train command line option --data-path ilab-model-train command line option --deepspeed-cpu-offload-optimizer ilab-model-train command line option --deepspeed-cpu-offload-optimizer-pin-memory ilab-model-train command line option --deepspeed-cpu-offload-optimizer-ratio ilab-model-train command line option --device ilab-model-train command line option --effective-batch-size ilab-model-train command line option --enable-serving-output ilab-model-evaluate command line option --endpoint-url ilab-data-generate command line option ilab-model-chat command line option --few-shots ilab-model-evaluate command line option --filename ilab-model-download command line option --gguf-model-path ilab-model-train command line option --gpu-layers ilab-model-serve command line option --gpus ilab-model-evaluate command line option ilab-model-train command line option --greedy-mode ilab-model-chat command line option --hf-token ilab-model-download command line option --input-dir ilab-model-train command line option --interactive ilab-config-init command line option --is-padding-free ilab-model-train command line option --iters ilab-model-train command line option --judge-backend ilab-model-evaluate command line option --judge-model ilab-model-evaluate command line option --learning-rate ilab-model-train command line option --legacy ilab-model-train command line option --local ilab-model-train command line option --log-file ilab-model-serve command line option --lora-alpha ilab-model-train command line option --lora-dropout ilab-model-train command line option --lora-quantize-dtype ilab-model-train command line option --lora-rank ilab-model-train command line option --lora-target-modules ilab-model-train command line option --max-batch-len ilab-model-train command line option --max-ctx-size ilab-model-serve command line option --max-seq-len ilab-model-train command line option --max-tokens ilab-model-chat command line option --max-workers ilab-model-evaluate command line option --merge-system-user-message ilab-model-evaluate command line option --min-taxonomy ilab-config-init command line option --model ilab-data-generate command line option ilab-model-chat command line option ilab-model-evaluate command line option ilab-model-test command line option --model-dir ilab-model-convert command line option ilab-model-download command line option ilab-model-test command line option

--model-family ilab-data-generate command line option ilab-model-chat command line option ilab-model-serve command line option ilab-model-test command line option --model-name ilab-model-convert command line option --model-path ilab-config-init command line option ilab-model-serve command line option ilab-model-train command line option --nnodes ilab-model-train command line option --node-rank ilab-model-train command line option --non-interactive ilab-config-init command line option --num-cpus ilab-data-generate command line option --num-epochs ilab-model-train command line option --num-instructions ilab-data-generate command line option --num-threads ilab-model-serve command line option --output-dir ilab-data-generate command line option ilab-model-evaluate command line option --quick-question ilab-model-chat command line option --quiet ilab-data-generate command line option ilab-taxonomy-diff command line option --rdzv-endpoint ilab-model-train command line option --rdzv-id ilab-model-train command line option --release ilab-model-download command line option --repository ilab-config-init command line option ilab-model-download command line option --rouge-threshold ilab-data-generate command line option --save-samples ilab-model-train command line option --server-ctx-size ilab-data-generate command line option --session ilab-model-chat command line option --skip-de-quantize ilab-model-convert command line option --skip-preprocessing ilab-model-train command line option --skip-quantize ilab-model-convert command line option ilab-model-train command line option --tasks-dir ilab-model-evaluate command line option --taxonomy-base ilab-config-init command line option ilab-data-generate command line option ilab-taxonomy-diff command line option --taxonomy-path ilab-config-init command line option ilab-data-generate command line option ilab-model-evaluate command line option ilab-taxonomy-diff command line option --test_file ilab-model-test command line option --tls-client-cert ilab-data-generate command line option ilab-model-chat command line option ilab-model-evaluate command line option --tls-client-key ilab-data-generate command line option ilab-model-chat command line option ilab-model-evaluate command line option --tls-client-passwd ilab-data-generate command line option ilab-model-chat command line option ilab-model-evaluate command line option --tls-insecure ilab-data-generate command line option ilab-model-chat command line option ilab-model-evaluate command line option --tokenizer-dir ilab-model-train command line option --train-profile ilab-config-init command line option --verbose ilab command line option --version ilab command line option --warmup-steps ilab-model-train command line option --yaml-rules ilab-data-generate command line option ilab-taxonomy-diff command line option -c ilab-model-chat command line option -gm ilab-model-chat command line option -m ilab-model-chat command line option ilab-model-test command line option -qq ilab-model-chat command line option -s ilab-model-chat command line option -sd ilab-model-convert command line option -sq ilab-model-convert command line option ilab-model-train command line option -t ilab-model-test command line option -v ilab command line option

A

additional_args (instructlab.configuration._train attribute)

api_base() (instructlab.configuration._serve method)

B

backend (instructlab.configuration._serve attribute) base_branch (instructlab.configuration._evaluate attribute)

base_model (instructlab.configuration._evaluate attribute) batch_size (instructlab.configuration._mmlu attribute) branch (instructlab.configuration._evaluate attribute)

C

chat (instructlab.configuration.Config attribute) chat_template (instructlab.configuration._serve attribute) chunk_word_count (instructlab.configuration._generate attribute)

ckpt_output_dir (instructlab.configuration._train attribute) context (instructlab.configuration._chat attribute) CRITICAL (instructlab.configuration.LogLevel attribute)

D

data_output_dir (instructlab.configuration._train attribute) data_path (instructlab.configuration._train attribute)

DEBUG (instructlab.configuration.LogLevel attribute) debug_level (instructlab.configuration._general attribute) deepspeed_cpu_offload_optimizer (instructlab.configuration._train attribute)

E

effective_batch_size (instructlab.configuration._train attribute)

ERROR (instructlab.configuration.LogLevel attribute) evaluate (instructlab.configuration.Config attribute)

F

few_shots (instructlab.configuration._mmlu attribute)

G

general (instructlab.configuration.Config attribute) generate (instructlab.configuration.Config attribute)

gpu_layers (instructlab.configuration._serve_llama_cpp attribute) gpus (instructlab.configuration._evaluate attribute) greedy_mode (instructlab.configuration._chat attribute)

H

host_port (instructlab.configuration._serve attribute)

I

ilab command line option --config --verbose --version -v ilab-config-init command line option --interactive --min-taxonomy --model-path --non-interactive --repository --taxonomy-base --taxonomy-path --train-profile ilab-data-generate command line option --api-key --chunk-word-count --endpoint-url --model --model-family --num-cpus --num-instructions --output-dir --quiet --rouge-threshold --server-ctx-size --taxonomy-base --taxonomy-path --tls-client-cert --tls-client-key --tls-client-passwd --tls-insecure --yaml-rules ilab-model-chat command line option --api-key --context --endpoint-url --greedy-mode --max-tokens --model --model-family --quick-question --session --tls-client-cert --tls-client-key --tls-client-passwd --tls-insecure -c -gm -m -qq -s QUESTION ilab-model-convert command line option --adapter-file --model-dir --model-name --skip-de-quantize --skip-quantize -sd -sq ilab-model-download command line option --filename --hf-token --model-dir --release --repository ilab-model-evaluate command line option --backend --base-branch --base-model --batch-size --benchmark --branch --enable-serving-output --few-shots --gpus --judge-backend --judge-model --max-workers --merge-system-user-message --model --output-dir --tasks-dir --taxonomy-path --tls-client-cert --tls-client-key --tls-client-passwd --tls-insecure

ilab-model-serve command line option --backend --chat-template --gpu-layers --log-file --max-ctx-size --model-family --model-path --num-threads ilab-model-test command line option --adapter-file --api-key --data-dir --model --model-dir --model-family --test_file -m -t ilab-model-train command line option --ckpt-output-dir --data-output-dir --data-path --deepspeed-cpu-offload-optimizer --deepspeed-cpu-offload-optimizer-pin-memory --deepspeed-cpu-offload-optimizer-ratio --device --effective-batch-size --gguf-model-path --gpus --input-dir --is-padding-free --iters --learning-rate --legacy --local --lora-alpha --lora-dropout --lora-quantize-dtype --lora-rank --lora-target-modules --max-batch-len --max-seq-len --model-path --nnodes --node-rank --num-epochs --rdzv-endpoint --rdzv-id --save-samples --skip-preprocessing --skip-quantize --tokenizer-dir --warmup-steps -sq ilab-taxonomy-diff command line option --quiet --taxonomy-base --taxonomy-path --yaml-rules INFO (instructlab.configuration.LogLevel attribute) instructlab.configuration module is_padding_free (instructlab.configuration._train attribute)

J

judge_model (instructlab.configuration._mtbench attribute)

L

llama_cpp (instructlab.configuration._serve attribute) llm_family (instructlab.configuration._serve_llama_cpp attribute) (instructlab.configuration._serve_vllm attribute)

log_level (instructlab.configuration._general attribute) logs_dir (instructlab.configuration._chat attribute) lora_quantize_dtype (instructlab.configuration._train attribute) lora_rank (instructlab.configuration._train attribute)

M

max_batch_len (instructlab.configuration._train attribute) max_ctx_size (instructlab.configuration._serve_llama_cpp attribute) max_seq_len (instructlab.configuration._train attribute) max_tokens (instructlab.configuration._chat attribute) max_workers (instructlab.configuration._mtbench attribute) mmlu (instructlab.configuration._evaluate attribute) mmlu_branch (instructlab.configuration._evaluate attribute) model (instructlab.configuration._chat attribute) (instructlab.configuration._evaluate attribute) (instructlab.configuration._generate attribute)

model_path (instructlab.configuration._serve attribute) (instructlab.configuration._train attribute) module instructlab.configuration mt_bench (instructlab.configuration._evaluate attribute) mt_bench_branch (instructlab.configuration._evaluate attribute)

N

nproc_per_node (instructlab.configuration._train attribute) num_cpus (instructlab.configuration._generate attribute)

num_epochs (instructlab.configuration._train attribute) num_instructions (instructlab.configuration._generate attribute)

O

output_dir (instructlab.configuration._generate attribute) (instructlab.configuration._mtbench attribute)

P

prompt_file (instructlab.configuration._generate attribute)

Q

QUESTION ilab-model-chat command line option

S

save_samples (instructlab.configuration._train attribute) seed_file (instructlab.configuration._generate attribute)

serve (instructlab.configuration.Config attribute) session (instructlab.configuration._chat attribute)

T

tasks_dir (instructlab.configuration._mmlubranch attribute) taxonomy_base (instructlab.configuration._generate attribute)

taxonomy_path (instructlab.configuration._generate attribute) (instructlab.configuration._mtbenchbranch attribute) train (instructlab.configuration.Config attribute)

V

vi_mode (instructlab.configuration._chat attribute) visible_overflow (instructlab.configuration._chat attribute)

vllm (instructlab.configuration._serve attribute) vllm_args (instructlab.configuration._serve_vllm attribute)

W

WARNING (instructlab.configuration.LogLevel attribute)