daslab-testing
/

CloverLM

Text Generation

low-precision-training

Model card Files Files and versions

mansaripo commited on 14 days ago

Commit

43e4bf9

·

verified ·

1 Parent(s): 389961d

Update README.md

Files changed (1) hide show

README.md +4 -2

README.md CHANGED Viewed

@@ -121,8 +121,10 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
 model = AutoModelForCausalLM.from_pretrained(
     "daslab-testing/CloverLM",
     trust_remote_code=True,
-    torch_dtype="bfloat16",
-)
 tokenizer = AutoTokenizer.from_pretrained(
     "daslab-testing/CloverLM",
     trust_remote_code=True,

 model = AutoModelForCausalLM.from_pretrained(
     "daslab-testing/CloverLM",
     trust_remote_code=True,
+    dtype="bfloat16",
+    quartet_2_impl="quartet2",  # native NVFP4 kernel or "pseudoquant" on non-Blackwell GPUs
+).to("cuda")  # for GPU usage or "cpu" for CPU usage
 tokenizer = AutoTokenizer.from_pretrained(
     "daslab-testing/CloverLM",
     trust_remote_code=True,