docs(inference): update deepseek-r1-distill-llama-70b:fp8 max context length

docs(inference): update deepseek-r1-distill-llama-70b:fp8 max context length #13024