Added gemma specific fp8 quantization file

huggingface · Oct 21, 2024 · 2376b58 · 2376b58
1 parent c625680
commit 2376b58
Showing 1 changed file with 12 additions and 0 deletions.
diff --git a/examples/text-generation/quantization_config/maxabs_quant_gemma.json b/examples/text-generation/quantization_config/maxabs_quant_gemma.json
@@ -0,0 +1,12 @@
+{
+    "method": "HOOKS",
+    "mode": "QUANTIZE",
+    "observer": "maxabs",
+    "scale_method": "maxabs_hw",
+    "blocklist": {"types": [], "names":  [
+        "matmul_qk",
+        "matmul_av",
+        "lm_head"
+    ]},
+    "dump_stats_path": "./hqt_output/measure"
+}