Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

#0: Update Llama3 PERF.md and llama3 vision PCC after cache regeneration #17505

Merged
merged 1 commit into from
Feb 3, 2025
Merged

#0: Update Llama3 PERF.md and llama3 vision PCC after cache regeneration #17505

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 11 additions & 11 deletions models/demos/llama3/PERF.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,15 +10,15 @@ This configuration uses bfp4 MLP FF1+FF3 for all models.

| Model | Device | Top-1 (%) | Top-5 (%) | Speed (t/s/u) |
|-------|--------|-----------|-----------|---------------|
| 1b | N150 | 90 | 98 | 91.0 |
| 1b | N300 | 90 | 98 | 98.8 |
| 1b | T3K | 88 | 98 | 97.8 |
| 1b | N150 | 87 | 98 | 91.0 |
| 1b | N300 | 87 | 98 | 98.8 |
| 1b | T3K | 87 | 98 | 97.8 |
| 1b | TG | 88 | 99 | 51.0 |
| 3b | N150 | 91 | 98 | 49.2 |
| 3b | N150 | 90 | 98 | 49.2 |
| 3b | N300 | 90 | 98 | 56.8 |
| 3b | T3K | 91 | 98 | 54.5 |
| 3b | T3K | 88 | 98 | 54.5 |
| 3b | TG | 90 | 97 | 33.5 |
| 8b | N150 | 88 | 99 | 28.6 |
| 8b | N150 | 86 | 99 | 28.6 |
| 8b | N300 | 85 | 98 | 38.9 |
| 8b | T3K | 84 | 97 | 53.7 |
| 8b | TG | 86 | 98 | 29.5 |
Expand All @@ -36,16 +36,16 @@ This configuration uses bfp4 MLP FF1+FF3 only for the 3.1-70B model.
| Model | Device | Top-1 (%) | Top-5 (%) | Speed (t/s/u) |
|-------|--------|-----------|-----------|---------------|
| 1b | N150 | 89 | 98 | 86.8 |
| 1b | N300 | 88 | 100 | 98.1 |
| 1b | T3K | 90 | 99 | 97.5 |
| 1b | N300 | 88 | 99 | 98.1 |
| 1b | T3K | 86 | 99 | 97.5 |
| 1b | TG | 87 | 98 | 51.3 |
| 3b | N150 | 92 | 100 | 44.2 |
| 3b | N300 | 92 | 99 | 54.2 |
| 3b | T3K | 90 | 99 | 55.6 |
| 3b | T3K | 91 | 98 | 55.6 |
| 3b | TG | 91 | 98 | 33.6 |
| 8b | N150 | 91 | 99 | 23.6 |
| 8b | N300 | 92 | 99 | 34.5 |
| 8b | T3K | 91 | 99 | 49.8 |
| 8b | N300 | 91 | 99 | 34.5 |
| 8b | T3K | 90 | 99 | 49.8 |
| 8b | TG | 88 | 100 | 29.5 |
| 11b | N300 | 91 | 99 | 33.8 |
| 11b | T3K | 91 | 99 | 52.6 |
Expand Down
2 changes: 1 addition & 1 deletion models/demos/llama3/tests/multimodal/test_llama_cross_attention_transformer_text.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,7 +56,7 @@ def test_llama_cross_attention_transformer_text_inference(
):
dtype = ttnn.bfloat8_b
prefill_pcc_required = 0.98
decode_pcc_required = 0.97
decode_pcc_required = 0.965

mesh_device.enable_async(True)

Expand Down
Loading