Hugging Face's logo

newmindai
/
Llama-3.1-8B-Instruct-w16a16-4nodes-bs32

Text Generation
Transformers
Safetensors
PyTorch
Turkish
llama
legal
turkish
instruction-tuned
llama-3.1
fp8
bfloat16
mixed-precision
question-answering
fsdp-v2
distributed-training
torchao
conversational
text-generation-inference
Model card Files
xet
Community
nmmursit commited on
Commit
5bd50d3
·
verified ·
1 Parent(s): 5e7eae5

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -4
README.md CHANGED
@@ -99,13 +99,14 @@ model = fully_shard(model, **fsdp_kwargs)
99
  | tensorboard-data-server | 0.7.2 |
100
  | wandb | 0.22.1 |
101
 
102
-
103
-
104
  ## Job Details
105
  | model | Job ID | Runtime (mins) | Nodes | GPUs | Node-hour | GPU-hour | micro-batch | batch-size | gradient_accumulation | total_batch_size |
106
  | ---------------------------------------- | -------- | -------------- | ----- | ---- | --------- | ---------- | ----------- | ---------- | --------------------- | ---------------- |
107
- | Llama-3.1-8B-Instruct-w16a16-tw | 31472940 | 51.50 | 1 | 4 | **0.858** | **3.433** | 2 | 2 | 4 | 32 |
108
- | Llama-3.1-8B-Instruct-w16a8-1node-bs8 | 31473092 | 47.25 | 1 | 4 | **0.788** | **3.151** | 2 | 2 | 4 | 32 |
 
 
 
109
  | Llama-3.1-8B-Instruct-w16a16-4nodes-bs32 | 31478433 | 31.75 | 4 | 4 | **2.117** | **8.467** | 4 | 4 | 8 | 512 |
110
  | Llama-3.1-8B-Instruct-w16a8-4nodes-bs32 | 31478468 | 39.75 | 4 | 4 | **2.650** | **10.600** | 4 | 4 | 8 | 512 |
111
  | Llama-3.1-8B-Instruct-w16a16-8nodes-bs32 | 31476914 | 22.00 | 8 | 4 | **2.933** | **11.733** | 4 | 4 | 8 | 1024 |
 
99
  | tensorboard-data-server | 0.7.2 |
100
  | wandb | 0.22.1 |
101
 
 
 
102
  ## Job Details
103
  | model | Job ID | Runtime (mins) | Nodes | GPUs | Node-hour | GPU-hour | micro-batch | batch-size | gradient_accumulation | total_batch_size |
104
  | ---------------------------------------- | -------- | -------------- | ----- | ---- | --------- | ---------- | ----------- | ---------- | --------------------- | ---------------- |
105
+ | Llama-3.1-8B-Instruct_w16a8_rw | 31768103 | 115.75 | 1 | 4 | **1.929** | **7.716** | 2 | 2 | 4 | 32 |
106
+ | Llama-3.1-8B-Instruct_w16a8_rw_with_gw_hp| 31837629 | 109.00 | 1 | 4 | **1.816** | **7.266** | 2 | 2 | 4 | 32 |
107
+ | Llama-3.1-8B-Instruct-w16a8-mxtw | 31768031 | 64.00 | 4 | 4 | **1.066** | **4.266** | 2 | 2 | 4 | 32 |
108
+ | Llama-3.1-8B-Instruct-w16a16-tw | 31768074 | 138.75 | 1 | 4 | **0.858** | **3.433** | 2 | 2 | 4 | 32 |
109
+ | Llama-3.1-8B-Instruct-w16a8-1node-bs8 | 31768093 | 123.75 | 1 | 4 | **0.788** | **3.151** | 2 | 2 | 4 | 32 |
110
  | Llama-3.1-8B-Instruct-w16a16-4nodes-bs32 | 31478433 | 31.75 | 4 | 4 | **2.117** | **8.467** | 4 | 4 | 8 | 512 |
111
  | Llama-3.1-8B-Instruct-w16a8-4nodes-bs32 | 31478468 | 39.75 | 4 | 4 | **2.650** | **10.600** | 4 | 4 | 8 | 512 |
112
  | Llama-3.1-8B-Instruct-w16a16-8nodes-bs32 | 31476914 | 22.00 | 8 | 4 | **2.933** | **11.733** | 4 | 4 | 8 | 1024 |