rename jamba example (#1846) [skip ci]

* rename jamba example

* feat: change readme

---------

Co-authored-by: Gal Cohen <galc@ai21.com>
This commit is contained in:
Gal Cohen (galco)
2024-08-22 16:22:55 +03:00
committed by GitHub
parent f07802f9fa
commit 957c956f89
3 changed files with 47 additions and 36 deletions

View File

@@ -6,5 +6,5 @@
- ✅ qlora w/ deepspeed Zero-3 needs at least 2x GPUs and 67GiB VRAM (wtf?)
- ✅ qlora single-gpu, ~51GiB VRAM
- ✅ multipack
- FSDP
- FSDP
- ❓ 8-bit LoRA

View File

@@ -1,4 +1,4 @@
base_model: ai21labs/Jamba-v0.1
base_model: ai21labs/AI21-Jamba-1.5-Large
tokenizer_type: AutoTokenizer
load_in_4bit: true
@@ -11,7 +11,7 @@ datasets:
drop_system_message: true
dataset_prepared_path: last_run_prepared
val_set_size: 0.0
output_dir: jamba-fsdp-qlora-ft
output_dir: jamba-large-fsdp-qlora-ft
save_safetensors: true
adapter: qlora
sequence_len: 2048