Update documentation

Signed-off-by: Pierangelo Di Pilato <[email protected]>

Author: pierDipi

deploy/config/sim-epp-kvcache-config.yaml

@@ -15,7 +15,8 @@ plugins:
         blockSize: 16                         # must match vLLM block size if not default (16)
         hashSeed: "42"                        # must match PYTHONHASHSEED in vLLM pods
       tokenizersPoolConfig:
-        tokenizersCacheDir: "/cache/tokenizers"
+        hf:
+          tokenizersCacheDir: "/cache/tokenizers"
       kvBlockIndexConfig:
         enableMetrics: false                  # enable kv-block index metrics (prometheus)
         metricsLoggingInterval: 6000000000   # log kv-block metrics as well (1m in nanoseconds)

docs/architecture.md

@@ -330,7 +330,8 @@ plugins:
           blockSize: 64
           hashSeed: "12345"
       tokenizersPoolConfig:
-        huggingFaceToken: your_hf_token_here    # automatically set by `HF_TOKEN` environment variable
+        hf:
+          huggingFaceToken: your_hf_token_here    # automatically set by `HF_TOKEN` environment variable
       kvBlockIndexConfig:
         enableMetrics: true
 ```
@@ -359,8 +360,9 @@ plugins:
             enableMetrics: true
           tokenizersPoolConfig:
             workersCount: 8
-            huggingFaceToken: your_hf_token_here    # automatically set by `HF_TOKEN` environment variable
-            tokenizersCacheDir: /tmp/tokenizers
+            hf:
+              huggingFaceToken: your_hf_token_here    # automatically set by `HF_TOKEN` environment variable
+              tokenizersCacheDir: /tmp/tokenizers
 ```
 
 ---