Merge pull request #2958 from AI-Hypercomputer:shuningjin-xlml

PiperOrigin-RevId: 857306849

Author: Google-ML-Automation

end_to_end/tpu/mixtral/8x7b/2_test_mixtral.sh

@@ -41,7 +41,7 @@ python3 -m MaxText.decode "${MAXTEXT_PKG_DIR:-${MAXTEXT_REPO_ROOT:-$PWD}/src/Max
 python3 -m MaxText.decode "${MAXTEXT_PKG_DIR:-${MAXTEXT_REPO_ROOT:-$PWD}/src/MaxText}/"configs/base.yml load_parameters_path=${UNSCANNED_CKPT_PATH} run_name=unscanned_decoding per_device_batch_size=1 model_name=mixtral-8x7b async_checkpointing=false tokenizer_path="${MAXTEXT_ASSETS_ROOT:-${MAXTEXT_PKG_DIR:-${MAXTEXT_REPO_ROOT:-$PWD}/src/MaxText/assets}}"/tokenizer.mistral-v1 ici_tensor_parallelism=1 ici_fsdp_parallelism=-1 max_prefill_predict_length=11 max_target_length=24 prompt='"[INST] I love to [/INST]"' megablox=False sparse_matmul=False scan_layers=false capacity_factor=1.25
 
 # Test whether the forward pass logits match the golden logits - matmul implementation
-python3 -m tests.forward_pass_logit_checker "${MAXTEXT_PKG_DIR:-${MAXTEXT_REPO_ROOT:-$PWD}/src/MaxText}/"configs/base.yml base_output_directory=${BASE_OUTPUT_PATH} load_parameters_path=${UNSCANNED_CKPT_PATH} run_name=matmul_forward_pass_test per_device_batch_size=1 model_name=mixtral-8x7b tokenizer_path="${MAXTEXT_ASSETS_ROOT:-${MAXTEXT_PKG_DIR:-${MAXTEXT_REPO_ROOT:-$PWD}/src/MaxText/assets}}"/tokenizer.mistral-v1 ici_tensor_parallelism=1 ici_fsdp_parallelism=-1 max_prefill_predict_length=11 max_target_length=11 dtype=float32 megablox=False sparse_matmul=False scan_layers=false --token_size=4 --max_kl_div=2e-3
+python3 -m tests.forward_pass_logit_checker "${MAXTEXT_PKG_DIR:-${MAXTEXT_REPO_ROOT:-$PWD}/src/MaxText}/"configs/base.yml base_output_directory=${BASE_OUTPUT_PATH} load_parameters_path=${UNSCANNED_CKPT_PATH} run_name=matmul_forward_pass_test per_device_batch_size=1 model_name=mixtral-8x7b tokenizer_path="${MAXTEXT_ASSETS_ROOT:-${MAXTEXT_PKG_DIR:-${MAXTEXT_REPO_ROOT:-$PWD}/src/MaxText/assets}}"/tokenizer.mistral-v1 ici_tensor_parallelism=1 ici_fsdp_parallelism=-1 max_prefill_predict_length=11 max_target_length=11 dtype=float32 megablox=False sparse_matmul=False scan_layers=false --token_size=4 --max_kl_div=3e-3
 
 # To repeat duplicate tests, we have MoE unit test to verify outputs matching for matmul, megablox, and ragged_dot implementation at https://github.com/AI-Hypercomputer/maxtext/blob/5c4090b8d5713a1a25cab85df89b0ec9c9862635/MaxText/tests/moe_test.py#L338-L411