Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
@@ -38,13 +38,13 @@ accross various devices, can be found [here](https://aihub.qualcomm.com/models/l
|
|
38 |
- Token generator output: 1 output token + KV cache outputs
|
39 |
- Use: Initiate conversation with prompt-processor and then token generator for subsequent iterations.
|
40 |
|
41 |
-
| Model | Device | Chipset | Target Runtime | Response Rate (Tokens/Second) | Time To First Token
|
42 |
|---|---|---|---|---|---|
|
43 |
-
| Llama-v3-8B-Chat | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 66.14 | (0.028, 0.92) | -- | -- |
|
44 |
-
| Llama-v3-8B-Chat | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 66.14 | (0.028, 0.92) | -- | -- |
|
45 |
-
| Llama-v3-8B-Chat | Samsung Galaxy S23 Ultra | Snapdragon® 8 Gen 2 | QNN | 66.14 | (0.028, 0.92) | -- | -- |
|
46 |
-
| Llama-v3-8B-Chat | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 66.14 | (0.028, 0.92) | -- | -- |
|
47 |
-
| Llama-v3-8B-Chat | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 66.14 | (0.028, 0.92) | -- | -- |
|
48 |
|
49 |
## Deploying Llama 3 on-device
|
50 |
Please follow [this tutorial](https://github.com/quic/ai-hub-apps/tree/main/tutorials/llama)
|
|
|
38 |
- Token generator output: 1 output token + KV cache outputs
|
39 |
- Use: Initiate conversation with prompt-processor and then token generator for subsequent iterations.
|
40 |
|
41 |
+
| Model | Device | Chipset | Target Runtime | Response Rate (Tokens/Second) | Time To First Token Range (Seconds)
|
42 |
|---|---|---|---|---|---|
|
43 |
+
| Llama-v3-8B-Chat | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 66.14 | 0.028 - 0.92 | (0.028, 0.92) | -- | -- |
|
44 |
+
| Llama-v3-8B-Chat | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 66.14 | 0.028 - 0.92 | (0.028, 0.92) | -- | -- |
|
45 |
+
| Llama-v3-8B-Chat | Samsung Galaxy S23 Ultra | Snapdragon® 8 Gen 2 | QNN | 66.14 | 0.028 - 0.92 | (0.028, 0.92) | -- | -- |
|
46 |
+
| Llama-v3-8B-Chat | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 66.14 | 0.028 - 0.92 | (0.028, 0.92) | -- | -- |
|
47 |
+
| Llama-v3-8B-Chat | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 66.14 | 0.028 - 0.92 | (0.028, 0.92) | -- | -- |
|
48 |
|
49 |
## Deploying Llama 3 on-device
|
50 |
Please follow [this tutorial](https://github.com/quic/ai-hub-apps/tree/main/tutorials/llama)
|