Hugging Face's logo

ibm-ai-platform
/
llama3-8b-accelerator

Transformers
Safetensors
mlp_speculator
Model card Files
xet
Community
4
JRosenkranz commited on
Commit
e723a51
·
verified ·
1 Parent(s): 6a039c0

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -3
README.md CHANGED
@@ -118,7 +118,7 @@ pip install transformers==4.35.0 sentencepiece numpy
118
  ```bash
119
  MODEL_PATH=/path/to/llama3/hf/Meta-Llama-3-8B-Instruct
120
  python fms-extras/scripts/paged_speculative_inference.py \
121
- --variant=8b \
122
  --model_path=$MODEL_PATH \
123
  --model_source=hf \
124
  --tokenizer=$MODEL_PATH \
@@ -135,7 +135,7 @@ python fms-extras/scripts/paged_speculative_inference.py \
135
  ```bash
136
  MODEL_PATH=/path/to/llama3/hf/Meta-Llama-3-8B-Instruct
137
  python fms-extras/scripts/paged_speculative_inference.py \
138
- --variant=8b \
139
  --model_path=$MODEL_PATH \
140
  --model_source=hf \
141
  --tokenizer=$MODEL_PATH \
@@ -151,7 +151,7 @@ python fms-extras/scripts/paged_speculative_inference.py \
151
  ```bash
152
  MODEL_PATH=/path/to/llama3/hf/Meta-Llama-3-8B-Instruct
153
  python fms-extras/scripts/paged_speculative_inference.py \
154
- --variant=8b \
155
  --model_path=$MODEL_PATH \
156
  --model_source=hf \
157
  --tokenizer=$MODEL_PATH \
 
118
  ```bash
119
  MODEL_PATH=/path/to/llama3/hf/Meta-Llama-3-8B-Instruct
120
  python fms-extras/scripts/paged_speculative_inference.py \
121
+ --variant=llama3.8b \
122
  --model_path=$MODEL_PATH \
123
  --model_source=hf \
124
  --tokenizer=$MODEL_PATH \
 
135
  ```bash
136
  MODEL_PATH=/path/to/llama3/hf/Meta-Llama-3-8B-Instruct
137
  python fms-extras/scripts/paged_speculative_inference.py \
138
+ --variant=llama3.8b \
139
  --model_path=$MODEL_PATH \
140
  --model_source=hf \
141
  --tokenizer=$MODEL_PATH \
 
151
  ```bash
152
  MODEL_PATH=/path/to/llama3/hf/Meta-Llama-3-8B-Instruct
153
  python fms-extras/scripts/paged_speculative_inference.py \
154
+ --variant=llama3.8b \
155
  --model_path=$MODEL_PATH \
156
  --model_source=hf \
157
  --tokenizer=$MODEL_PATH \