#Instructlab Release v0.21.0 https://github.com/instructlab/instructlab/releases/tag/v0.21.0
This release includes various enhancements and stability improvements to the ilab workflow:
- System Profiles & System Profile Auto-Detection
- Full training pipeline memory optimizations
- Granite architecture models (starting with granite-3.0-8b)
- ilab data generate speed optimizations using --max-num-tokens
- A switch to using Granite-7b as the default serve/chat model
#llm #genai #aiml
Markus Eisele /
npub1l9…qn8d6
2024-11-20 20:14:00
Author Public Key
npub1l95d5gnp63dy5t4l43h8lpztmlrd3ej77tpnsjsj2hmk88ey5ftq9qn8d6Published at
2024-11-20 20:14:00Event JSON
{
"id": "0db780607a56e9a71b28a449751a2677a94a5d7eca88761c97fc17cfd08afa38",
"pubkey": "f968da2261d45a4a2ebfac6e7f844bdfc6d8e65ef2c3384a1255f7639f24a256",
"created_at": 1732133640,
"kind": 1,
"tags": [
[
"t",
"instructlab"
],
[
"t",
"llm"
],
[
"t",
"genai"
],
[
"t",
"AIML"
],
[
"proxy",
"https://mastodon.online/users/myfear/statuses/113517110259361431",
"activitypub"
]
],
"content": "#Instructlab Release v0.21.0 https://github.com/instructlab/instructlab/releases/tag/v0.21.0\n\nThis release includes various enhancements and stability improvements to the ilab workflow:\n- System Profiles \u0026 System Profile Auto-Detection\n- Full training pipeline memory optimizations\n- Granite architecture models (starting with granite-3.0-8b)\n- ilab data generate speed optimizations using --max-num-tokens\n- A switch to using Granite-7b as the default serve/chat model \n#llm #genai #aiml",
"sig": "d1e9ce7b37d881e9cd5f29a8170495052de0cf40583dd88cab3074baf5ec9a60157ebfa39dc60a898fb08d1c3f13453fff425ef4186c6945c85b909b87e29c88"
}