[Neuron] Auto-select StaticCache when device is Neuron

[dacorvo]

Context

On Neuron devices, StaticCache is required for correct generation — dynamic tensor shapes trigger per-step recompilations. Currently, users must explicitly pass past_key_values=StaticCache(...) or cache_implementation="static" to model.generate(). Ideally, model.generate() should auto-select StaticCache when running on Neuron, providing zero-config DX.

Depends on: #44742 (StaticCache-friendly _sample must exist first — otherwise auto-selecting StaticCache still hits the dynamic-shape _sample path)

Proposed change

In _prepare_cache_for_generation (around line 1879 of src/transformers/generation/utils.py), when no cache is specified:

if generation_config.cache_implementation is None and self.device.type == "neuron":
    generation_config.cache_implementation = "static"

Precedent: The generation code already has device-aware logic:

• Line 2021: self.device.type in ["cuda", "xpu"] for compile criteria
• Line 1870: Forces "dynamic_full" for assisted generation

Device-aware cache selection fits the existing pattern.

Expected behavior

# On Neuron — auto-selects StaticCache, uses static-shape _sample path:
model.generate(input_ids)

# On CUDA — unchanged, uses DynamicCache + standard _sample:
model.generate(input_ids)

# Explicit override still works on any device:
model.generate(input_ids, cache_implementation="static")
model.generate(input_ids, past_key_values=my_static_cache)

Related

• Parent: [Neuron] Improve transformers compatibility with AWS Neuron devices #44741
• Depends on: [Neuron] Static-shape generation loop for compilation-friendly inference #44742